Frank Summers - Object Relations Theories and Psychopathology - A Comprehensive Text-The Analytic Press (1994)

O b je c t R e l a t io n s T h e o r ie s
and P sych o path o lo g y

This page intentionally left blank
FRANK SUMMERS
O b je c t R e l a t io n s T h e o r ie s
and P sych o pa th o lo g y
A Comprehensive Text
U J Psychology Press
A Taylor &. Francis Group
NEW YORK AND LONDON
First published 1994 by
The Analytic Press
Published 2014 by Psychology Press

711 Third Avenue, New York, NY 10017
and by Psychology Press

27 Church Road, Hove, East Sussex, BN3 2FA
Psychology Press is an imprint o f the Taylor & Francis Group,

an informa business
Copyright © 1994 by The Analytic Press
All rights reserved. No part of this book may be reprinted or reproduced or utilised
in any form or by any electronic, mechanical, or other means, now known or
hereafter invented, including photocopying and recording, or in any information
storage or retrieval system, without permission in writing from the publishers.
Trademark notice: Product or corporate names may be trademarks or registered

trademarks, and are used only for identification and explanation without intent to
infringe.
Library of Congress Cataloging in Publication Data
Summers, Frank
Object relations theories and psychopathology : a comprehensive text
Frank Summers,
p. cm.
Includes bibliographical references and index.
1. Object relations (Psychoanalysis) I. Title.
1DNLM: 1. Object Attachment. 2. Psychopathology. WM 460.5.02
S955o 19941
RC455.4.023S85 1994
616.89’ 001 - - dc20 93-39264
ISBN 13: 978-0-881-63155-5 (hbk)

To Renee
Contents
Acknowledgments ix
Preface xi
1 - The Origins of Object Relations Theories 1
2 - The Work of W. R. D. Fairbairn and Harry Guntrip 25
3 - The Work of Melanie Klein 73
4 - The Work of D. W. Winnicott 137
5 - The Work of Otto Kernberg 191
6 - The Work of Heinz Kohut 247
7 - The Interpersonalists 311
8 - An Object Relations Paradigm for Psychoanalysis 345
References 381
Index 391
vii
Acknowledgments
I wish to express my gratitude to Dr. Paul Stepansky for the excep

tional effort he put into editing this manuscript. His helpful sugges
tions improved the manuscript with respect to both content and style.
I also wish to thank Toby Troffkin for a quality job of copy editing and
Eleanor Starke Kobrin for her assistance. I want to give special thanks
to my students over the years in the Division of Psychology, North
western University Medical School, for their stimulating, thought-
provoking discussion, and persistent encouragement, without which
this book would never have come to pass.
Preface
T h is b o o k is in t e n d e d f o r b o t h b e g in n in g a n d e x p e r ie n c e d c l in ic ia n s
who may be interested in either learning about object relations theo
ries for the first time or integrating these theoretical approaches more
fully into their work. While the contributions of many different theo
reticians are worthy of inclusion in a text of this type, this book con
centrates on object relations theorists who are most clinically
relevant. For this reason, such highly regarded and influential theo
rists as Margaret Mahler and Edith Jacobson, although discussed, are
not subjects of intensive investigation. It is hoped that the reader will
gain an appreciation for the way each theoretician (and his or her fol
lowers) deliberates upon and uses clinical material to further the
therapeutic process. Only in the last chapter is an effort made to
move toward an integration of these diverse clinical approaches, and
even then there is no attempt to blur critical clinical distinctions. Each
chapter is an invitation to the reader to step into the theoretical
worldview of the theorist and approach clinical material without
uncritical acceptance of it.
A word about the chapter sequence is in order. Although the chap
ters are roughly chronological, a strictly temporal sequence was not
adopted. Although Klein's writings antedate those of Fairbairn, the
chapter discussing her work follows that devoted to Fairbairn and
Guntrip. Although many of the significant writings of Klein and Fair
bairn were contemporaneous and each theorist influenced the other,
Fairbairn's work is more easily grasped and represents a cleaner
break with classical analytic theory. For these reasons, his work
serves better as an introduction to object relations theorizing than
does Klein's writing. In addition, although Kohut's major theorizing
began before Kernberg's primary contributions, the continuity of the
text is best maintained by placing Kernberg's work, as well as Winni-
cott's, directly after the discussion of Klein because both theorists
were influenced by her whereas Kohut was not. The chapter on the
interpersonal theorists follows the chapters on the object relations
xi
xii Preface
theorists to enrich the reader's appreciation of the contributions of

this related school of thought and to provide deeper understanding
by contrasting the two theoretical approaches.
This book aims to give the reader a comprehensive understanding
of the major object relations theories. Each theory is presented from
the viewpoint of those primary assumptions and principles out of
which its ideas of pathology and treatment grow. Further, and per
haps most important, the emphasis here is on the clinical process.
Each theory has critical implications for the conduct of psychoana
lytic treatment, implications that grow organically from its view of
development and pathology. In the discussion of each theory, the
emphasis is on what its principles mean for the practicing clinician.
CHAPTER 1
The Origins of
Object Relations Theories
O b je c t r e l a t io n s t h e o r ie s h a v e b e e n w i d e l y r e a d a n d d is c u s s e d in
recent years as psychoanalytic theorists, and clinicians who have

begun to question traditional psychoanalytic theory have turned
increasingly to object relations theories to broaden or even supplant
their theoretical and clinical understanding. Nonetheless, there is a
great deal of confusion regarding the nature of many object relational
theories and their clinical application. Object relations theories differ
widely with respect to key concepts, assumptions, and principles and
are often confusing and difficult to digest.
In addition, the general trend of object relations theories has been
subject to widely different interpretations. Greenberg and Mitchell
(1983) view object relations theories as part of a wider movement to
supplant drive theory with a relational model of psychoanalysis;
Kernberg (1984) and Winnicott (1960a) tend to see object relations con
cepts as an addition to drive theory applicable to more primitive emo
tional disorders; and Bacal and Newman (1990) view object relations
theories as a bridge to self psychology. An object relations theory as
defined here and used throughout the text signifies any systematic
effort to account for personality development and pathology on the
basis of the internalization of relationships with others. This model is
contrasted with the drive-ego model, according to which the drives
and their vicissitudes (however disguised, sublimated, or neu
tralized), along with ego mechanisms, account for personality
development.
In an attempt to consolidate knowledge of object relations theories,
Greenberg and Mitchell (1983) presented the first comprehensive
review of each of the major theories. Serving the critical purpose of
familiarizing clinicians with the major concepts of these theories, their
text has probably made the single most significant contribution to the
dissemination of information about object relations. Despite the
invaluable contributions of their discussion of object relations theo
ries, however, there are two key drawbacks to their work. First,
1
2
Greenberg and Mitchell do not present each theory as an integrated

whole; they discuss the various ideas of each theory in isolation with
out full appreciation of the unifying principles on which the individ
ual ideas are based. Thus, the reader fails to obtain a sense of the
overall thrust of each theory. Second, Greenberg and Mitchell pursue
the discussion of each theory primarily in terms of its proximity to
either the drive /structure or the relational/structure model. While
this issue is pivotal theoretically, their focus on thus categorizing each
theory gives short shrift to its clinical implications. Nor is it clear from
their theoretical discussions how drive /structure and relational/
structure models differ clinically.
Since the publication of their book, both Mitchell (1988) and
Greenberg (1991) have offered clinical theories that involve relational
concepts. These theories, which will be discussed in chapter 7, do not,
however, address the clinical implications of the major object relations
theories.
The contributions of the various object relations theories to the psy
choanalytic process remain unclear. This obfuscation may be due
partly to the cumbersome and difficult language often used by theo
rists. A large share of the difficulty, however, is due to the fact that
object relations theories were developed by clinicians dissatisfied with
the clinical and theoretical limitations of the classical psychoanalytic
models. We will see in our discussion of the various theorists that they
differ on whether their ideas are an addition to the classical viewpoint
or a replacement of it, and at times, the theorist is unclear about the
relationship between his or her ideas and the classical model. Each
object relations theory breaks away from the major tenets of the model
of endogenous drives to some degree, but, because of this confusion,
the clinical implications of this theoretical shift are not readily appar
ent. It is a major task of this book to show the connection between
theory and clinical intervention in each major object relations theory.
Because each object relations theory is a reaction to the classical the
ory, one cannot grasp the meaning and importance of its concepts
without an appreciation of the history of the psychoanalytic ideas
which preceded them. Object relations theories developed because
each theorist found some limitation in the drive-ego model that
pushed his or her thinking to new concepts and, ultimately, ways of
practicing psychoanalytic therapy. Consequently, to understand the
contributions of the object relational theorists, one must first grasp the
major trends in psychoanalytic theory. Therefore, we shall review in
detail the major developments in ego psychology so as to clarify the
ego-psychological view of psychic structure, development, pathology,
Origins of Object Relations Theories 3
and treatment. From this conceptual clarification, we can see the

growth of object relations theories from their roots in ego psychology
to the critical step of separation from it in the development of new
theorizing. As we will see, some object relations theorists take this
step more dramatically than others—hence the differences among
them in their degree of adherence to the classical model.
This chapter describes the shift from drive to ego psychology and
presents the major theoretical and clinical developments within the
ego psychology tradition. It highlights the major theoretical bridges
from ego psychology to object relations theories. This discussion
sets the context for the presentation of object relations theories in
subsequent chapters.
THE ORIGINS OF EGO PSYCHOLOGY
When Freud abandoned the theory of sexual trauma as the etiology of

neurosis in favor of endogenous drives, psychoanalytic theory shifted
to consideration of the internal workings of the mind. Freud (1915a)
began to focus on inborn drives as the motivating force of psy
chopathology and eventually extended this focus to personality devel
opment in general. As drives and their vicissitudes came to be
considered the critical factor in development, psychoanalytic theory
made a decisive move away from external events, including trauma,
toward the functioning of the mind, now conceived as a product of the
drives, or biological tension states, that aim for gratification through
tension reduction. According to this drive model, human motivation
originates in the press of biological drives that gain psychological
expression in the form of wishes that power psychological functioning.
Psychopathology, in this model, is caused by the repression of wishes,
not of memories of external trauma (Freud, 1915b). The pathogenic
conflict is between preconscious censorship of conscious thoughts and
unconscious wishes for instinctual gratification (Freud, 1915c). When
the repression barrier is broken through by disguised expressions of
unconscious wishes, symptoms result. The clinical implication of this
shift from the trauma theory to the drive model is that the goal of the
analytic process becomes the uncovering of unconscious wishes, the
repression of which is considered the cause of neurosis.
The study of unconscious wishes and their manifestation in psy
chopathology dominated psychoanalytic theory and practice from
1897 to 1923. This situation began to change when Freud (1923)
pointed out in The Ego and the Id that the unconscious cannot be
equated with wishes, nor the conscious with the forces of repression
4 Chapter One
because the mechanism of repression is unconscious. These facts led

Freud to superimpose his structural model of ego-id-superego upon
the topographic model of unconscious-preconscious-conscious: the
ego, consisting of repressive mechanisms, is largely unconscious; and
the superego, which is the moral system motivating the repression of
unacceptable wishes, has a conscious component, the conscience, as
well as an unconscious component in the form of unconscious guilt.
From the viewpoint of this structural model, psychological conflict
takes place not between the unconscious and the conscious, but
between the unconscious components of the ego or superego and the
id. Psychopathology is a compromise formation between an id con
tent, such as a sexual wish, that is otherwise blocked from conscious
ness, owing to its unacceptability to the moral system, or superego,
and an ego defense mechanism, such as repression. Psychopathology
is, therefore, a result of conflict between competing psychological
structures. This shift in Freud's theoretical thinking marked a change
in psychoanalytic theory and practice away from the exclusive focus
on drives to an equal emphasis on the forces opposing it (Freud, 1937).
With this theoretical shift, the ego now assumed the central role in
the functioning of the psyche. The degree of health or pathology of the
personality, from this viewpoint, is a function of the ability of the ego
to manage the press of drive-based wishes for discharge as well as the
constraints of reality on such gratification. The ego must also change
the moral constraints from within (in the form of the superego), which
constitutes an additional counterpressure to drive discharge. Ego
strength, or the capacity of the ego to handle the conflicting demands
of id, reality, and superego, now assumes the pivotal role in the well
being of the personality. To the degree that the ego is not able to
accomplish a functional balance, the personality will fall ill. For exam
ple, if the ego is forced to use excess repression, wishes will seek sub
stitute expressions of discharge and hysterical symptoms will result.
To the extent that the ego displaces unacceptable wishes onto the
environment, phobic fears ensue. Thus does ego psychology include
the functioning of the ego in all psychopathology, with every symp
tom implying a failure of the ego to balance effectively the need for
drive discharge with the constraints of superego strictures and reality
(Fenichel, 1945).
Freud (1926) changed his concept of anxiety in accordance with the
structural model. Whereas he had originally viewed anxiety as the
result of dammed up libido due to repression, he now reconceptual
ized anxiety as a warning signal to the ego. When the ego senses dan
ger from unacceptable wishes, it experiences anxiety; it then employs
Origins of Object Relations Theories
a defense mechanism to ward off the threatening affect and restore

balance, sometimes at the price of a symptom outbreak. From the
viewpoint of the structural model, anxiety is not a product of repres
sion, rather, it motivates repression and other defenses. This reconcep
tualization of anxiety reflects the central role of the ego in balancing
the various pressures to which the psyche is subject.
The structural model resulted in the concept of psychological orga
nization: the ego is not simply a group of mechanisms but a coherent
organization whose task is to master the competing pressures of the
id, superego, and relations with reality. This concept led Freud to
question how such mastery is possible given the biological origins of
the psyche. He had to account for the establishment of a psychological
organization, the structured ego, that opposes the gratification of the
drives, from which all human motivation originates.
Freud's answer was that the ego develops from drive frustration.
Simple drive gratification is never all the child wants, even in the best
of circumstances; eventually, the preoedipal tie to the mother is given
up, and later the oedipal object must be relinquished (Freud, 1923).
The loss of these early objects, according to Freud, forces the child to
set up a substitute: an internal psychological representation of the par
ents to replace the abandoned objects of childhood longings. As the
early attachment to the mother is given up in reality, she is taken in
psychologically. The object cathexis of the mother is replaced by iden
tification with her. In the oedipal phase, which Freud considered
decisive for identifications, the longing for the parent of the opposite
sex is given up and the child either intensifies identification with the
same sex parent or identifies with the opposite sex parent in response
to the loss. These identifications will determine the gender identifica
tion of the personality and concomitantly form the superego-ego ideal
complex. Each relinquishment of a childhood object results in an
identification that helps form the ego structure:
When it happens that a person has to give up a sexual object, there quite
often ensues an alteration of his ego which can only be described as a
setting up of the object inside the ego. . . .the process, especially in the
early phases of development, is a very frequent one, and it makes it
possible to suppose that the character of the ego is a precipitate of aban
doned object-cathexes and that it contains the history of those object-
choices [Freud, 1Q23, p. 29].
In Freud's formulation, the id drives the organism to seek object

contact to achieve instinctual gratification. When reality forces the
6 Chapter One
relinquishment of these objects, they are taken in through identifica

tion and form the basis of the ego. Thus, the ego develops out of the
frustration of id wishes and is formed by becoming like the objects
reality forces the id to relinquish. Likewise, the superego is a "precipi
tate of abandoned object-cathexes of the id," but it is also a reaction
formation against those choices in the form of moral objection. Thus,
both the ego and the superego are formed from the internalization of
previously cathected objects.
THE CLASSICAL EGO PSYCHOLOGISTS
Freud's pioneering suggestions regarding the importance of the ego

and the mechanisms of its development became the basis for ego psy
chology, which extended the concept of the ego even further than
Freud did. Anna Freud (1936) enumerated a variety of defensive
mechanisms used by the ego to keep wishes unconscious. She
pointed out that the various defenses used by the ego become resis
tances in the analytic process; thus, psychoanalytic treatment is
focused equally on ego mechanisms and id wishes. She drew further
implications from the structural model by pointing out that psycho
analytic assessment of development and psychopathology must
include the functioning of the ego.
Subsequent ego psychologists have extended further the concept of
the ego's autonomy from the id. Hartmann (1939) pointed out that
some of the mechanisms used by the ego, such as perception, motility,
and memory, do not develop from frustration but are autonomously
developed functions, which he termed "apparatuses of primary ego
autonomy," which later become integrated and are necessary for the
functioning of the ego. Hartmann pointed out that since these ego
functions exist from birth and originate outside of conflict, one cannot
properly speak of the ego as developing "out of" the id; rather, both
ego and id gradually develop from an undifferentiated matrix and
become separate systems. This concept removes the original depen
dence of the ego on the id that characterized Freud's formulation.
Hartmann referred to that part of ego functioning which is not in con
flict at any given time as the "conflict-free ego sphere." For example,
while fantasy is at times a product of frustrated wishes and conflict, it
is also a useful means for the consideration of alternatives in solving
problems.
The concept of autonomous ego apparatuses does not mean that
Hartmann disputed Freud's view that drive frustration leads to the
structuralization of the ego. Indeed, in Hartmann's view there are
two sources of ego development: the motivation of inborn appara

tuses of primary autonomy and the frustration of drives, which results
in secondary autonomy. With regard to the second source, Hartmann
agreed with Freud that energy from libidinal frustration is used for
the organized ego; however, in his view aggression, rather than
libido, is a more significant factor in ego structuralization. (Hartmann,
Kris, and Lowenstein, 1949). Hartmann pointed out that since the
intent of the aggressive drive is to destroy the object, its discharge is
more dangerous than that of the libidinal drive and, consequently, its
neutralization is more criticaL For the same reason, in the view of
Hartmann and his colleagues, permanent object relations are more
dependent on the sublimation of the aggressive than of the libidinal
drive. Neutralized aggression leads to structuralization of the ego,
which allows for good object relationships and object constancy and
which, in turn, make possible the further neutralization of aggression.
Unneutralized aggression, on the other hand, is accorded a primary
role in much of psychopathology (for example, when unneutralized
aggression attacks an organ, psychosomatic illness results). According
to Hartmann (1953) when aggression is not neutralized, no counter-
cathexis is possible, aggression erupts over the organism, object
relations are not possible, and a schizophrenic process results.
In Hartmann's view, the ego is a group of functions, including
defenses and adaptive mechanisms. These functions are organized
into a system Hartmann called the "synthetic function" of the ego.
This system is not simply an outgrowth of the id but an organized,
adaptive capacity that controls healthy functioning and has its own
sources of growth in addition to frustration of wishes. Nonetheless,
complete ego autonomy is not possible, in Hartmann's view, because
the ego uses energy from the drives, especially the aggressive drive.
Thus, the organized ego is always linked to the id and achieves only
relative autonomy from it.
Rapaport (1951,1957) viewed the id as a constitutionally given and
the ego as the created personality. A lthough he agreed with
Hartmann that the ego develops from an undifferentiated ego-id
matrix, Rapaport pointed out that in healthy development the emer
gent ego organization obeys its own laws, distinct from and indepen
dent of the elements from which it emerged. To the extent that the ego
is independent of the id, it is better adapted to reality and more capa
ble of functioning; the extent to which the ego is unable to achieve
autonomy from the id is the degree to which it will be a slave to it,
with a resultant inability to adapt to the demands of reality. The
health of the personality, in Rapaport's view, is a function of ego
8
autonomy, that is, the ability of the ego to manage id pressures. The
actual content of id wishes and the conflicts to which they give rise are
of little moment to Rapaport, as the same wishes and conflicts may
exist in healthy and pathological personalities; the difference lies in
the ability of the healthy ego to achieve autonomy from the id so that
it can manage its conflicts without symptomatic outcome.
Arlow and Brenner (1964) extended Hartmann's concept of ego
autonomy by pointing out that the topographic model was not modi
fied by the structural model but replaced by it. In contrast to the com
mon ego-psychological view, as represented by Hartmann and Anna
Freud, that the topographic model can be used along with the struc
tural model, the view held by Arlow and Brenner is that the two mod
els are, in fact, contradictory because anti-instinctual forces are
unconscious. With the introduction of the structural model, conflicts
were no longer considered to be between the preconscious and uncon
scious; both instinctual wish and the force that opposes it are seen as
unconscious. The concept of the preconscious had been obviated by
the central role of the ego in psychic conflict; because the preconscious
could not determine the nature of the psychic content with which the
instinctual wish is in conflict, it was replaced by the ego.
In agreement with Rapaport, Brenner (1981) endorses Hartmann's
view that the ego and id develop from an undifferentiated matrix. He
points out, however, that since all mental phenomena include some
degree of compromise between ego and id, the two are not separable
except under conditions of conflict. The ego as executant of the id
must find a way to help it achieve instinctual gratification. To accom
plish this goal, the ego must negotiate the dangers to which all id
wishes give rise. Therefore, according to Brenner (1976), mental phe
nomena are products of a compromise formation including wish,
guilt, anxiety, and defense. The id wish conflicts with feelings of guilt,
creating anxiety that is warded off by defense. The task for the ego is
to find a way to allow instinctual gratification within the limits set by
guilt feelings and anxiety.
To perform this task, the ego uses a variety of mental mechanisms,
such as fantasy, perception, cognition, and the functions typically
labeled "defense mechanisms" (Brenner, 1981). In this view of mental
functioning, defenses are not a specialized group of mechanisms, as
conceptualized by Anna Freud (1936) and Hartmann (1939); that is,
one cannot label any particular ego function as a "defense" since all
ego functions have both defensive and adaptive value (Brenner, 1981).
Healthy, socially acceptable behavior is no less a compromise forma
tion than is a symptom. When instinctual gratification is excessively
compromised to satisfy the demands of guilt and anxiety, symptoms

or pathological character traits result. The decisive factor in health or
illness for Brenner (1976) and Arlow (1963) is the ability of the ego to
execute a compromise formation that allows instinctual gratification
without symptoms.
White (1963) took the final step in the theoretical movement
toward the concept of the autonomous ego with his view that the
ego has its own independent energies. Unlike other ego psycholo
gists, White based his position on animal and child development
research. He pointed out that there is abundant evidence from ani
mal research to support the concept of nonbiological motivation: a
variety of animals will learn mazes and solve problems when all
their drive needs are satisfied. Further, rats and other animals that
have been studied will learn complex material solely for the experi
ence of novelty and the opportunity to explore, and among rewards
of novelty they prefer objects they can manipulate and have an effect
on. White noted that Harlow's monkeys would learn solely for the
reward of looking outside and that the learning curve for this experi
ence was similar to that achieved for biological rewards. White con
cluded that there is a drive for mastery over the environment, which
he termed "effectance motivation."
White believed that experimental and naturalistic observation of
infants and young children supported the notion of the existence of a
need for "effectance" independent of biological motivation. He noted
that observations of infants as early as the first few days after birth
show that they spend some time in exploration and that time for this
activity gradually increases until one-year-olds spend about six hours
a day in playful exploration. Infants perform activities during this
"playtime" for no reward other than the successful completion of the
behavior. White noted Piaget's observation that infants as young as
three months learn to repeat behavior for the sole purpose of having
an effect on the environment—and show clear signs of delight when
they are successful. White also pointed out that children sometimes
choose to perform activities that, in fact, delay gratification but lead to
the mastery of a skill. For example, children prefer to use a spoon
rather than their hands to eat, thus delaying instinctual gratification
but promoting the joy of mastery.
White concluded that the need for effectance is not only separate
from, but may also be in conflict with, biological drives. In White's
view, effectance motivation is fueled by energies inherent in the ego
apparatus that are totally independent of instinctual needs. Hendrick
(1942,1943), from a more purely psychoanalytic viewpoint, also came
10 Chapter One
to the conclusion that a drive for mastery of the environment fuels the
organization of the ego. Hendrick (1942) resolved the issue, however,
by postulating a drive "to do and learn to do" (p. 40). White pointed
out that the motivation to have an effect on the environment shares
none of the characteristics of a drive per se, that it has neither somatic
source nor consummatory pattern nor specificity of aim. White's
argument is that only independent ego energies can account for the
animal and child research data and that, therefore, the ego does not
develop from a common matrix with the id but is separate and
autonomous from birth. White's position is the final step in the evolu
tion of ego psychology toward the liberation of the ego from its
dependence on the id.
In W hite's formulation there is no role for abandoned object
cathexes in the development of the ego. Independent ego energies
require no object relations to achieve structuralization; the parents'
role in ego development is to provide identificatory objects for model
ing effectance. The child wants to be like the father in order to achieve
competence in affecting the environment. This concept of identifica
tion is very different from Freud's (1923) view of "taking the object in"
in order to withstand the pain of loss. White viewed identification
as a form of imitation, not as an incorporation of the object, as it
was for Freud.
Whether ego psychology views the ego as completely independent
of the id, as in White's formulation, or more functionally autonomous,
as in the theories of Hartmann and Rappaport, the ego is seen as a
separate organization from the id. This model of the mind, consisting
of drives and an ego organization that has some autonomous ability to
regulate their discharge is referred to here as the "drive-ego model."
From this viewpoint the crucial issues in development are the vicissi
tudes of the drives and the concomitant organization of the ego, the
adaptive capacity of the organism. As can be seen from this review,
within ego psychology there are two views of ego autonomy: the view
of Hartmann and Rapaport that the ego originates from both inborn
apparatuses and the neutralization of drives and the view of White
and Hendrick that the ego is formed solely from its own energies. For
Hartmann and Rapaport, who adhered to Freud's notion that the ego
is formed partly on the basis of drive frustration, ego autonomy is rel
ative. By contrast, White's more complete break with Freud led to an
abandonment of the frustration model of ego development and the
notion of structuralization through frustration. White was able to mar
shall considerable experimental and observational evidence to sup
port his view of total ego autonomy; consequently, he gave only a
minimal role to object relationships in the development of the ego.

White viewed the ego as originating in psychic energy, similar to the
energy fueling the id. By contrast, Hartmann and Rapaport saw in
their view of the relative autonomy of the ego, a direct connection
between the vicissitudes of object relations and ego development,
with frustration in drive-fueled object relationships leading to the
structuralization of the ego. Their derivation of the ego from id
energy was speculative, however, lacking the evidential support
of White's theory.
EGO PSYCHOLOGY AND THE PSYCHOANALYTIC PROCESS
The classical ego psychologists tended to emphasize theory, but they

did draw some clinical implications from their view of the importance
of the ego in development and pathology. Rapaport (1954) pointed out
that the clinician cannot make an assessment solely on the basis of
knowledge of the patient's drives and their vicissitudes. Such an
assessment would leave out of account the functioning of the personal
ity i its strengths and weaknesses and would not be sufficient to
determine the patient's prognosis and suitability for analysis. Owing to
the influence of ego psychology, psychoanalytically informed assess
ment now typically includes judgments regarding the patient's ego
strengths and weaknesses as well as its structure. Psychopathology is
understood not simply in terms of the conflicts that produced it but
also from the way the individual handles the conflict, that is, by the
ego mechanisms used for this purpose. From the ego-psychological
viewpoint, the structure of the ego is as necessary to understanding
pathology as is the conflict with which the ego is grappling.
With regard to the psychoanalytic process, it has been mentioned
that Anna Freud (1936) adopted the view, endorsed by her father
(Freud, 1940), that analysis is only half about unconscious v/ishes,
thp other half being concerned with the ego and superego, their
structure and functioning. Interpretation is geared toward the ego
mecnanisms as much as toward what they conceal. Although this
may seem like a self-evident technical principle from the contempo
rary viewpoint, it is a clear departure from the position Freud (1895)
adopted in Studies on Hysteria in which he advocated using any
means to circumvent the patient's defenses to bring forth repressed
material. Ego psychology shifted the theory of technique to defense
interpretation, according this aspect of the process a role equal to
that of interpreting unconscious wishes.
Arlow (1987) and Brenner (1976) extended the concept of defense
Chapter One
analysis by drawing out the clinical implications of the view that

symptoms are compromise formations effected by the ego. They point
out that defense analysis is not a clinical process distinguishable from
the analysis of wishes; that is, one does not analyze defenses first.
Since every symptom is a compromise formation, its analysis includes
interpretation of the wish, guilt, anxiety, and defense that compose it.
As the psychology of mental conflict, psychoanalysis is always con
cerned with the unconscious conflict between a wish pressing for
instinctual discharge and the danger situation that its gratification
would produce. The analysis of the danger situation that motivates
defense may be termed ego analysis, but it is inseparable from other
aspects of interpretation. Indeed, in Arlow's (1963, 1969) view, both
the id wish and the danger to which it gives rise are unconscious fan
tasies; thus, pathology is considered to be ultimately a product of
unconscious fantasies, one from the side of the id and the other
motivating the ego's defense.
In the view of both Arlow (1987) and Brenner (1976), the role of the
analyst is solely to interpret unconscious conflict. The analytic task,
therefore, is to understand psychic conflict and make conscious the
compromise formations to which it gives rise. The patient tries to
enlist the analyst to gratify the unconscious wish; by not complying,
the analyst facilitates its expression (Arlow and Brenner, 1990). For
Brenner (1979), there is no role for the development of an analytic
alliance, nor any noninterpretive behavior by the analyst to form a
relationship with the patient. If the patient pressures the analyst for
an alliance, the analyst's role is to interpret the conflict that underlies
that wish, not to comply with it. Brenner's (1979) contention is that
analysts in the presence of severe emotional symptoms, such as
excessive dependence, suicidal ideation, and depression, too quickly
forget the importance of analyzing conflict, that the success of the
analysis is a function of the analyst's ability to maintain the analytic
stance irrespective of the severity of the conflict.
Because conflict is ubiquitous, defenses will still operate and a new
compromise formation will be effected after all the elements of the
conflict are made conscious (Brenner, 1976). However, the defenses,
now less intense, will result in a new compromise formation that
allows greater control and integration by the ego and increased
instinctual gratification. According to Arlow (1987) and Brenner
(1976), the result of a beneficial analysis is not a change in the defenses
but a new compromise formation that allows instinctual gratification
without symptom formation.
Despite these clinical implications of ego psychology, another
group of ego psychologists believes that there has been a "develop

mental lag" between the advances in ego psychological metapsychol
ogy and the clinical theory derived from this theoretical shift (Gray,
1982). Their contention is that, despite widespread acceptance of ego
psychology as the metapsychological foundation of psychoanalysis,
practice has continued to employ an id model. Gray points out that
Freud's original technique of circumventing the defenses in whatever
way necessary still has undue influence on technique, although in
more subtle fashion than overt suggestion and manipulation. For
example, if the patient is angry at the analyst, interpreting this quickly
as an aggressive drive derivative from childhood bypasses the nega
tive transference and the defenses against it, thus increasing and per
haps preserving resistance. To the extent that the analyst relies on
quick interpretation of impulses, she is using id, rather than ego, psy
chology, in conducting the clinical process. If the analyst insists on
confronting impulses that the patient is trying to keep unconscious, he
will necessarily meet resistance, and it may appear that suggestion or
manipulation is in order. According to Gray, however, such interven
tions are not called for. The reason for the increased resistance and
apparent need for noninterpretive intervention is to be found in the
analyst's technique of bypassing the defenses.
Gray (1990) proposes, instead, a technical model based on the
recognition that the patient's symptoms are a product of the defensive
processes of the ego. According to this model, the analyst listens for
drive derivatives but does not intervene until the patient's ego uncon
sciously interferes with the flow of material. This resistance reflects
the defensive functioning of the ego, and attention is directed to it. In
opposition to Arlow and Brenner, Gray believes that interpretation is
best approached from the side of the defense. He assumes that the
patient's resistance is caused by a fantasy of danger if certain words
are spoken to the analyst. By directing attention to the immediate
resistance, this fantasy will be addressed and may itself be revealed as
a defense. The goal of this type of intervention is to increase the
patient's awareness of his unconscious ego rather than to bring
impulses into consciousness.
Gray's fundamental point is that a strict adherence to the principles
of ego psychology dictates a focus on ego functioning within the ana
lytic session as the best means for understanding the way the ego
defends and adapts. Because the purpose of interpretation is to help
the patient give up his defenses, he must be shown how they work in
the analytic relationship. In Gray's (1973) view, a major advantage of
this approach is that it leads the patient to become self-observing.
14 Chapter One
Gray points out that the patient is most likely to become self-analytic
if he sees his defenses operating on a moment-to-moment basis.
Gray (1987) has applied the same reasoning to the treatment of the
superego. He feels that analysts tend to overlook the analysis of the
superego because Freud was pessimistic regarding the subjection of
this psychic agency to analytic scrutiny. Freud (1920, 1926) saw the
superego as a form of resistance linked to the death instinct and,
therefore, unanalyzable. In Gray's view, the superego is an alienated
part of ego functioning and should be analyzed like any symptom,
with the goal of making it conscious. As the nature and origins of the
superego become conscious, it is brought under the control of the ego.
Whereas Freud believed the superego could be influenced mostly by
suggestion, Gray sees it as an analyzable portion of the ego that will
enhance ego functioning when made conscious.
Weiss (1971) also emphasizes the importance of here-and-now
defense analysis, but his views depart more radically from the classical
theory of analytic technique. Weiss argues that unconscious urges do
not become conscious owing to frustration caused by the neutrality and
abstinence of the analytic setting. If this were so, he argues, their erup
tion into consciousness would be disruptive rather than helpful. Weiss
points out that defenses are given up when the ego feels it is safe to do
so, indicating that the lifting of the defenses is under the unconscious
control of the ego. When the ego judges reality to be safe, it lifts the
defenses, and the ego defenses change from being a segregated portion
of the ego to an ego-syntonic control mechanism in harmony with the
rest of the ego. In the analytic setting, this means that the patient will
test the analyst to judge whether the analyst can safely endure the reve
lation of anxiety-provoking impulses. When the analyst is so judged,
unconscious impulses can become conscious and will then be subject to
ego regulation. According to Weiss, this process explains why making
the unconscious conscious is helpful rather than disruptive. Weiss con
tends that classical theory has remained within the outmoded frustra
tion theory and has therefore failed to appreciate the role of the ego in
the clinical process, a state of affairs metaphorically referred to as a
"developmental lag" between the metapsychology of the ego and clini
cal theory. The value of defense analysis, according to Weiss, is that it
changes the relationship of the defenses to the rest of the ego, a psychic
shift that makes possible the appearance of unconscious wishes.
Weiss (1988) has reported empirical evidence to substantiate his
claim regarding the operation of the analytic process. On the basis of
blind ratings of clinical protocols in a limited number of cases, his data
show that when repressed contents became conscious, the patients'
anxiety was lower and their experience more vivid. The classical the
ory would predict the opposite, but the data support Weiss's hypoth
esis that the unconscious becomes conscious when the patient feels
safe rather than disrupted. In addition, Weiss found that after the
patients' unconscious demands were frustrated by the analyst, the
patients tended to feel less anxious, bolder, and more relaxed. If the
patient tested the analyst to have the demands gratified, as predicted
by the classical model, the patient would be more anxious. This find
ing confirms Weiss's hypotheses that patients test to see if the analyst
is safe and that if the analyst shows he or she is safe by maintaining
neutrality, patients feel relief and be able to bring forth more material.
In a similar study Weiss reported that patients' intensity of experi
ence and insight was aided by interpretations that tended to discon-
firm unconscious beliefs and that patients who did better in analysis
tended to receive such interpretations. In Weiss's view, these results
explain why some interpretations work and others do not. If an inter
pretation tends to confirm a patient's anxiety-provoking unconscious
belief, say, that he is inadequate, the interpretation will not help, but if
the interpretation tends to disconfirm the belief, as, for example, by
indicating the belief is a fantasy, it will bring relief. Weiss uses these
findings to support his contention that the analytic process works by
defense interpretation and m aintenance of a neutral analytic
stance, both of which help the patient feel safe, and that this sense
of safety opens the patient to previously warded-off impulses,
resulting in analytic success.
This group of ego psychologists has applied the insights of ego psy
chology to a reconceptualization of analytic technique. The impor
tance of drive interpretation recedes in their model in favor of detailed
attention to the operation of the defensive functioning of the ego in
the analytic setting. According to this model, impulses need not be
directly addressed, but will emerge when the defenses are properly
interpreted and when the patient feels safe after having successfully
tested the analyst's neutrality. To the degree that the analyst is able to
make the setting safe by adherence to analytic neutrality and defense
interpretation, progress toward the analytic goals will be made. The
approach of this group is a clear and consistent application of the
structural model to the analytic process.
EGO PSYCHOLOGY AND OBJECT RELATIONSHIPS
All the ego psychologists discussed thus far have either ignored or
minimized the role of object relationships in the formation of the ego.
16 Chapter One
In contrast, the third branch of ego psychology views ego formation as

a function of object relationships. This viewpoint, which provides the
foundation of object relations theories, is based on Freud's (1923) con
cept of the ego as the "precipitate of abandoned object cathexes"
(p. 29). W. R. D. Fairbairn in Scotland and Melanie Klein in London,
working independently, both inferred from this statement that the ego
consists of internalized object relations. Fairbairn (see chapter 2)
endorsed the concept of the ego's autonomy from the drives but
pointed out that the growth of the ego is dependent on satisfactory
object relationships. His protege, Harry Guntrip, whose theoretical
and clinical contributions are also examined in the second chapter,
further developed the relationship between early object relationships
and the growth or arrest of the ego. Klein (see chapter 3) also believed
that ego development is a product of internalized object relationships.
Klein, unlike Fairbairn and Guntrip, contended that endogenous libid
inal and aggressive drives give rise to object relationships and that
from the earliest phase of infancy these form the basis of the ego. The
significance she accorded the drives sets her theoretical views apart
from those of other object relations theorists, but her theoretical
and clinical system was based on the concept of ego growth
through the internalization of objects. Klein's modifications of
psychoanalytic theory spawned a group of followers who adopted
the fundamentals of her conceptual scheme but revised certain
aspects of it (see chapter 3).
Klein's views were sharply criticized by Anna Freud (1927), who,
as we have seen, adopted the more traditional ego psychological posi
tion that the ego was formed from the frustration of drives. The result
was a split in British psychoanalysis between the "Kleinians" and the
followers of Anna Freud (Segal, 1980). Analysts who neither fully
accepted the Kleinian system nor rejected all of its postulates became
known as the British "middle school," or "independents" (Rayner,
1991). Fairbairn and Guntrip are often included in this group along
with Michael Balint and Donald Winnicott (Sutherland, 1980). Balint
(1968), like Fairbairn, emphasized the importance of the primary love
relationship as the foundation of ego development. Winnicott was
influenced by Klein's object relations theory, but his views regard
ing the relationship between early object relations and ego devel
opment emphasized the mother-child bond rather than drives.
Winnicott's theoretical and clinical views, the subject of chapter 4,
are the most comprehensive system of thought to come from the
British middle school.
Edith Jacobson and Margaret Mahler
In America Edith Jacobson (1964) was the first theoretician to link

object relationships and the building of ego structure. In her view,
drive development, ego maturation, and the growth of object relation
ships are all aspects of a unified developmental process. She agreed
with Freud that pleasure and unpleasure are the primary infantile
experiences, but she pointed out that in the earliest phase of infancy
the child cannot distinguish pleasure from its source. In this phase, the
child's fantasies of merger with the mother form the foundation of all
future object relationships. Jacobson's contention was that the concept
of the oral stage had to be expanded beyond feeding and oral erotism
to a whole range of experiences that cluster around oral gratification
and frustration.
In Jacobson's view, at about three months the maturation of the ego
leads the child to differentiate the love object from the self. At this
point the beginnings of self- and object-images are formed, and they
cluster along the lines of the drive organization. Gratifying experi
ences become libidinally organized self-object units distinct and inde
pendent from the aggressively organized self-object units born of
frustrating experiences. In this phase every experience of closeness or
gratification leads to the temporary experience of return to the early
merger state. These fantasies are incorporative, or introjective, in the
sense that the child wishes to become the mother. Jacobson viewed
introjection, in contrast to customary usage, as a primitive mechanism
of incorporating the object in fantasy mechanism whereby the object
becomes the self. Analogously, she viewed projection as the primitive
experience of ejection, whereby the self becomes the object. In this
view, the earliest identifications consist of the re-fusion of self- and
object-images and are not true ego identifications.
According to Jacobson, if the mother is able to "tune in" to the dis
charge pattern of the infant, sometime in the first year the infant
begins a more active form of primitive identification by imitating the
parent. This new behavior is a developmental step forward because
the infant is active and utilizes an ego mechanism, motor behavior.
Nonetheless, Jacobson pointed out that imitation is not a true ego
identification because it is founded on the magical fantasy of
becoming the mother rather than on the wish to be like the mother.
Imitation, in Jacobson's view, is merger through activity rather
than through physical contact.
In the second year the child learns to distinguish the features of the
love object, and the temporal sense develops. These two capacities
18 Chapter One
allow the child to be like the object without the fantasy of being the
object. At this point selective identification begins to replace fusion as
a true ego identification, and the child is thus able to differentiate
wishful and real self-images. The child's wishful self-image and the
identification with the idealized parent form the ego ideal; this
benign structure compensates for the lost fusion. Concurrently, the
negative self-image built from frustrating experiences, realistic
parental prohibitions, and the ideal self- and object-images combine to
form the superego.
Jacobson pointed out that unless the child admires qualities of the
parents, he cannot form a meaningful identification in which the ego
is modified to assume characteristics of the object. Although she
believed that all identification has a component of separation, and
therefore of aggression, her contention was that the formation of the
ego is dependent primarily on the mother's attunement to the infant's
discharge needs. This attunement is the basis for the libidinal object
relationship, which fosters the development of positive self- and
object-images. These internalized positive images are the basic units
out of which the healthy ego is formed. Aggressive object relation
ships are inevitable, but if the positive self-images and object-images
are strong, they keep frustration within manageable limits, preventing
excessive aggressiveness. In turn, the strong ego is better able to with
stand gratifying experiences without merging and to experience
frustration without returning to primitive identifications. The interre
lationships among drive discharge, ego maturation, and object rela
tions are complex and reciprocal from Jacobson's point of view. The
healthy formation of ego and superego is inseparable from maternal
attunement to the child's discharge needs and the satisfaction of the
resulting object relationship.
Jacobson, unlike traditional ego psychologists, gave a primary role
to the nature of the early object relationship in the formation of psy
chological structure. Her view of the importance of this relationship
emphasized both gratification and frustration of drive needs, and both
are included in her concept of ego identification. For Jacobson, the
crucial process in early development is the shift from the fantasy
world of being the mother to being like the mother. This gradual move
ment from fantasy to reality is made possible by the reciprocal influ
ences of object relationships and identification. Jacobson's view that
the ego develops in accordance with early object relationships and
resulting identifications extended the connection between ego devel
opment and object relationships well beyond Freud's concept of aban
doned object cathexes. In Jacobson's view, the nature of the real
relationship between mother and child is crucial, and the bond that
grows out of this relationship is central in ways that go beyond its
frustrating component.
Like Jacobson, Mahler believed the real relationship between
mother and child is the crucial factor in the development of psycho
logical structure. Mahler, Pine, and Bergman (1975) saw the birth of
the psychological self as the outcome of the separation-individuation
process. They define separation as awareness of separateness from the
primary object and individuation as the assumption of individual
characteristics. Although Mahler and her coworkers conceptualized
this process as intrapsychic and therefore not directly observable, they
believed it could be inferred from systematic observations of the
behavior of infants and young children with their mothers.
Mahler agreed with Jacobson that in early infancy self and object
cannot be differentiated. According to M ahler's developmental
scheme, after a brief "autistic" phase devoid of object contact, the
infant of one to six months is in a "symbiotic phase" in which all expe
rience is fantasized as part of the self. The separation-individuation
process begins with the child's emergence from symbiosis and lasts
until the onset of the oedipal phase at about 36 months. The first sub
phase, from six to ten months, is termed "differentiation" and is char
acterized by the child's first awareness of its difference from the
environment. From about 10 to 16 months, the child engages in
"practicing," moving away from the mother both physically and emo
tionally to explore the world. According to Mahler, in the "rap
prochement" phase at about 16 to 25 months, the child suddenly
seems to realize that there is danger in moving away and seems to
want to return to the earlier bond. Nonetheless, the child still needs to
separate, and this resistance to losing its gains results in an "ambi-
tend ency." Eventually, the child moves away again tow ard
independence, and one can infer the existence of internalized
emotional object constancy in the child at this phase.
In Mahler's view, the intrapsychic process of separation-individua
tion, if successful, results in the internalization of whole objects with
both good and bad qualities. If the process is disturbed at any point,
however, the ego's development wTill be impaired and either a pre-
oedipal form of pathology will result or, at minimum, oedipal devel
opment will be distorted. Thus, Mahler, like Jacobson, viewed object
relationships as an inherent part of ego development. Mahler saw the
bond formed between mother and child, and the child's ability to use
it, as the crucial component in the child's internalization process and
consequent ego development. Like Jacobson's theory, Mahler's view
20 Chapter One
of ego development extended the role of object relationships well

beyond the concept of frustration.
The clinical implications of Jacobson's and Mahler's revisions of
psychoanalytic theory extend psychoanalytic treatment to severe psy
chopathology. Jacobson's view of ego development led her to treat
depressives, borderline patients, and even psychotic persons by
analytic means. She did, however, acknowledge that the treatment of
preoedipal psychopathology requires modification of the strict inter
pretive stance. Pregenital fantasies may often be used for the treat
ment process rather than interpreted. For example, Jacobson (1954)
allowed the idealizing transferences of depressed patients to go on for
extended periods without interpretation because she believed such
patients were attempting to recover their lost ability to love through
"magic love" of the analyst. To interpret such a transference quickly is
to interfere with the patient's need to use the analyst in a way that can
ultimately lead to restored functioning. Jacobson's primary contribu
tion to technique was to show that analytic treatment could be suc
cessful with severely disturbed patients as long as the analyst is
willing to be more flexible regarding interpretation than classical tech
nique allows. Jacobson appeared to draw no clinical conclusions from
her theoretical views for the analytic treatment of neurotic patients
and, consequently, made no effort to modify the classical model for
the treatment of such patients.
Mahler (1971, 1972), too, derived from her own theory a reconcep
tualization of the treatment of severe pathology. She viewed severe
forms of childhood psychosis in terms of her developmental theory:
infantile autism is fixation at the autistic level of development, and
childhood schizophrenia (Mahler, 1952,1968) is a pathological fixation
at the symbiotic phase. Mahler's contribution in these areas is a new
conceptualization of these severe illnesses based on an empirically
derived developmental model.
Although these conceptualizations in themselves are highly origi
nal, Mahler (1971, 1972) made perhaps her most unique contribution
to the understanding of pathology in the application of her develop
mental model to the borderline syndrome. In Mahler's view, the
borderline patient is caught in the rapprochement conflicts of the 16-
to-25-month-old child. The borderline patient, like the child of this
age, wishes to cling but fears the loss of Ids fragile sense of self, wishes
to separate, but fears the dangers of moving away from the parental fig
ure. In Mahler's view, treatment of the borderline patient, who is fix
ated at the rapproachement crisis, should focus on the patient's inability
to resolve the separation-individuation process. However, Mahler
provided no specific recommendations for treatment technique.

Although Mahler's clinical theory tends to focus on preoedipal
pathology, she, unlike Jacobson, did believe that her developmental
model had implications for neurosis. In Mahler's (1975) view, "the
infantile neurosis becomes manifestly visible during the oedipal
phase, but may be shaped by the rapprochement crisis that precedes it"
(p. 332). Conflicts in negotiating the rapprochement subphase render
more difficult the successful resolution of the oedipal phase and
thereby contribute to a neurotic outcome. In particular, Mahler
thought that neurotic patients who oscillate between desires for
merger and defense against it suffer from unresolved conflicts in
the rapprochement subphase. While she believed that her delin
eation of preoedipal developmental phases had significant implica
tions for at least some neurotic patients, she did not construct a
detailed therapeutic model for use with such patients.
The work of Klein, Jacobson, and, to a lesser degree, Mahler exerted
a strong influence on Otto Kernberg (see chapter 5). Kernberg adopted
Jacobson's blend of drives, object relations, and ego structuralization
to create a developmental theory based on the internalization of drive-
based object relations. Kernberg, like Jacobson, views development as
a series of object relationships with increasing degrees of structural
ization and differentiation. Like Klein and Jacobson, he believes drives
fuel psychological structure, but in his view drives are inherently
object relations and the ego is formed from object relations units.
Heinz Kohut's work does not clearly bear the stamp of Jacobson's
theory, but her theoretical influence can be seen in his abandonment
of the concept of the ego in favor of that of the self. While other object
relations theorists use the concept of the self, Kohut was the most
explicit in substituting self-structuralization for ego formation as the
foundation of the developmental process. Kohut, like all other object
relations theorists, viewed early object relationships as the key to the
formation of psychological structure, but he conceived of the psyche
as a self structure rather than as an organization of ego mechanisms.
Kohut's clinical theory was influenced by Jacobson's principle that
transference idealization in disturbed patients must be allowed to con
tinue for an extended period uninterrupted by interpretation. Kohut
applied this concept to the analysis of narcissistic disorders, for whom
he believed that a protracted period of idealization of the analyst is a
necessary step for the eventual recovery of narcissistic balance. His
systematic views on the development, pathology, and treatment of the
self (see chapter 6) have spawned self psychology as a separate
"school" within psychoanalysis.
22 Chapter One
CONCLUSION
In Freud's view, abandoned objects form ego structure through the

process of identification. The concept of ego autonomy, first intro
duced by Hartmann, challenged the contention that frustrating object
relations alone motivate ego development. If the ego has its own
sources of motivation, its development is not fueled by frustration.
It is true that if the ego has some degree of autonomy but is partly
motivated by drive neutralization, then Freud's concept of ego devel
opment from abandoned objects can still play a role in the structural
ization of the psyche. It is unclear, however, how such a speculative
transformation of psychic energy comes about, and in any case White
showed that ego functioning occurs in the absence of drive frustration.
Furthermore, Jacobson and Mahler showed that the mother-child rela
tionship includes much more than frustration and that aspects of the
whole relationship are internalized to form the ego and superego
structures. Their work demonstrated that the concept of ego auton
omy is not in conflict with the view that ego formation is a prod
uct of object relationships. All the major object relations theorists
have departed from Freud in adopting the view that psychological
structure is a product of object relationships and not simply their
frustrating aspects.
From the viewpoint of classical ego psychology, then, there is a
conflict between object relations and ego autonomy because ego for
mation derives from the frustration of drives. Because White viewed
object relations as rooted in drive frustration, he could not admit them
into his theory of autonomous ego development. Although he pro
vided impressive evidence of learning that was not drive motivated,
subsequent evidence indicates even more decisively that the infant
and growing child seek object contact independent of the drives
almost from birth (Bowlby, 1969; Lichtenberg, 1983; Stern, 1985). For
example, neonates seek the gaze of the parent and even search for it if
it is not there; in addition, they differentiate the mother from other fig
ures very early. Animal research shows that a variety of animals
attach to maternal figures for no reward other than contact itself. More
evidence of this type will be presented in chapter 8; here it suffices to
observe that there is abundant evidence that contact with objects is
not reducible to more primary drives. White, who had no concept of
autonomously motivated object relations, could understand ego
development only in terms of psychic energy.
Once the assumption is abandoned that psychological structure
necessarily grows out of frustration, it is possible to view personality
formation as a matter of the internalization of autonomously moti

vated object relationships. From this viewpoint, object relations theo
ries become an extension of the concept of complete ego autonomy
and also of the notion held by Freud, Hartmann and Rapaport that
the ego develops from object relations, though shorn of its assump
tion of drive frustration. All object relations theories view the person
ality as a complex product of early object relationships. Different
object relations theorists accord the drives different roles in this
process, but none of these theorists links the frustration of drives to
the structuralization of the personality.
This notion of the roots of personality organization is easily con
fused with interpersonal theory. Indeed, the concept of relational ori
gins of the psyche has been carried even further than the object
relations view by the interpersonal theorists. Beginning with Sulli
van's (1953) interpersonal theory of the personality, the Sullivanians
have developed a theory of personality formation, psychopathology,
and psychoanalysis based on the principle that all psychological phe
nomena are interpersonal. Until recent years the Sullivanians have
remained outside mainstream psychoanalysis; however, with the pub
lication of Greenberg and Mitchell's (1983) book showing the relation
ship between Sullivan and both the interpersonalists and the object
relations theorists, the interpersonal viewpoint has been given greater
consideration within established psychoanalytic theory (see chapter
7). Despite the shared emphasis on the relationship between self and
other in object relations and interpersonal theory, the relative neglect
of the internalization process in favor of the interpersonal situation in
the latter distinguishes the two types of theory.
Psychoanalytic work within the object relations paradigm is based
on a group of theories that, although differing considerably, have as
their underlying commonality the view of development and pathol
ogy as products of the internalization of early interpersonal relation
ships. Consequently, the conceptualization of the psychoanalytic
process in the paradigm is of a treatment focused on the manifesta
tions of these internalizations in the form of object relationships. Each
object relations theory has a different view of the critical factors in
development and pathology and a distinct concept of the significant
ingredients of successful analytic treatment. As we shall see, some the
ories tend to accord the drives a major role, whereas other theories
abandon drive theory entirely. The theories also differ on the role of
the environment versus constitution in pathology and on the critical
environmental variables implicated in questions of health and pathol
ogy. In their conceptualizations of the treatment process, they differ
Chapter One
on both the role and content of interpretation and on the extent to

which other interventions are desirable. Consequently, we cannot
speak of a single object relations theory.
Object relations is an umbrella concept for any theory that derives
its principles of human motivation from the need for early relation
ships and consequently views the primary goal of psychoanalytic
treatment as modification of the object relationships that have grown
out of these early relationships. Precisely how the personality devel
ops from early relationships and what the implications of this are for
treatment is answered differently by each theory, and we now turn to
the individual object relations theories to see how each variant of this
model addresses these issues.
CHAPTER 2
The Work of W. R. D. Fairbairn

And Harry Guntrip
W. R. D. FAIRBAIRN
Basic Concepts of Fairbairn's Theory
W. R. D. F a ir b a ir n (1949,1951) so u g h t t o r e c o n c e p t u a l iz e p s y c h o a n -
alytic theory by recasting it as an object relations model of personality
development and psychopathology. His interests remained the theo
retical reconceptualization of the psychoanalytic theory of develop
ment, mental structure, and pathology until near the end of his life,
when he began to draw out the clinical implications of his views.
Unfortunately, owing to his ill health and premature death, Fairbairn
was never able to complete his clinical theory, and it was left to his
analysand and protege Harry Guntrip to provide the clinical drama
for Fairbairn's object relations theory. Indeed, Fairbairn published no
case studies demonstrating his theoretical views. The theoretical focus
of Fairbairn's detailed reconceptualization of psychoanalytic theory,
the clinical application of which is often difficult to discern, gives his
writing a dry, abstruse quality. Nonetheless, there is a great deal of
theoretical innovation in Fairbairn's work, and Guntrip's writings
provide much of its concrete clinical application.
Fairbairn viewed as ground-breaking discoveries Freud's concept
of the unconscious and his interpretation of psychopathological phe
nomena and dreams as products of unconscious mental processes.
However, Fairbairn felt the limitation of Freud's thought lay in his
"impulse psychology," or drive theory, a limitation that led him to
formulate his alternative object relations psychology. For Fairbairn,
impulses, whether conscious or unconscious, exist only within an ego
structure, however primitive or undifferentiated it may be, and derive
their relevance from this. They do not somehow exist prior to the
development of the ego, either temporally or logically. For Fairbairn,
human experience can have meaning only in terms of an ego.
Consequently, he disputed Freud's (1923) concept of an original id out
25
26 Chapter 2
of which the ego is born as well as Hartmann's (1939) concept of an

undifferentiated matrix from which both ego and id develop. For
Fairbairn, the ego exists from birth. The baby's needs exist within an
experiencing organism, however undifferentiated it may be, a view
point that led Fairbairn to a primary theoretical postulate, namely,
that structure exists before energy. Fairbairn (1946) disagreed with
Freud's concept of a directionless psychic energy existing from birth
that must be harnessed as psychic structure develops; for him, the
concept of psychic energy is meaningful only insofar as it is associated
with an ego structure. Fairbairn ultimately disputed Freud's tripartite
structural model of id, ego, and superego, viewing all divisions in the
psyche as parts of the ego.
According to Fairbairn (1951), not only do all impulses emanate
from an experiencing ego but they always have objects. In this regard
he was influenced by the work of Melanie Klein. As will be discussed
in detail in chapter 3, Klein (1952b) believed that ego growth is a
process of internalizing objects. Fairbairn used Klein's concept of the
internal object as the building block of the personality to develop his
own object relations theory of development and psychopathology,
believing that Klein's object relations theory was too dependent on the
drive concept. In Fairbairn's (1949, 1951) view, libido is "object seek
ing." The infant cannot exist without an object, and, indeed, objects
are needed throughout life, although the type of need and the nature
of the relationship with the object changes. Infantile libido, like all
stages of libido, is conceptualized as object seeking. Thus, Fairbairn
reverses the primacy of libidinal zone and object. Freud (1905a) and
Abraham (1924) held to the erotogenic zone theory of development,
dividing the epigenetic stages of child development into the oral, anal,
and genital libidinal zones. In opposition to this view, Fairbairn pointed
out that the infant is oral not because of the primacy of the mouth, but
because the mouth is the appropriate organ for the breast. The child
becomes genital when it is able to have a more mature form of object
relationship. In other words, the erotogenic zone the child uses is
defined by the kind of object relationship it seeks and is capable of.
Throughout his work Fairbairn (1946) used the word libido to refer
to positive affective charge. However, he criticized as a hypostatiza-
tion Freud's notion of libido as existing originally in a directionless,
"pure" state. Fairbairn's (1944) view that libido exists within an ego
seeking an object from the beginning caused him to dispute the
importance of the pleasure principle in normal development and to
postulate that libido is reality oriented, not pleasure seeking, from the
start, although the relationship with reality is initially immature.
Fairbairn and Guntrip 27
Pleasure, according to Fairbairn, is a "signpost" to the object. When

pleasure is sought for its own sake, the psyche has broken down; a
personality dominated by pleasure seeking is pathological, even in the
earliest stages of development.
Fairbairn's defense of this view is that there is no other way to
understand the devotion of children and adults to their objects. Freud
(1920) recognized the problem of reconciling object attachment with
the pleasure principle when he raised the question of why neurotics
are so attached to painful objects; he resolved this dilemma with his
concept of the death instinct. Fairbairn's view, which does not require
the postulation of an abstract, unverifiable concept like the death
instinct, is that neurotics are attached to their bad objects because they
need them for survival, that any object contact is better than none at
all. If the organism were pleasure seeking, Fairbairn reasoned, the
internalization of bad objects would not be explainable, nor would the
attachment to objects no longer valued as sources of pleasure.
Fairbairn's rejection of the concept of the id in favor of the concept
of an ego present from birth did not mean he disputed Freud's notion
of psychic structures in conflict with each other. Instead, Fairbairn
(1944) reconceptualized Freud's structural model into three types of
ego, each of which has a corresponding object. In developing this con
ceptualization, Fairbairn was again influenced by Klein (1952b), who,
as we shall see in the next chapter, conceived of the early division of
the psyche as splits in the ego. In Fairbairn's theory, the dynamic of
psychic division is repression, as it is in Freud's. However, the nature
and developmental origins of psychic differentiation are quite differ
ent for the two theorists. To understand Fairbairn's conceptualization
of psychic structure, one must understand his developmental model,
and it is to this aspect of his reconceptualization of psychoanalytic
theory that we now turn.
Development
Fairbairn was influenced by Klein's (1957) concept of ego development

as a series of phases of object relationships. Like Klein, Fairbairn (1940)
redefined developmental stages in terms of the characteristic manner
of relating to objects. However, he opposed Klein's retention of the
concept that object relations are a function of the drive organization.
In Fairbairn's (1944) view, the earliest stage, in which orality is so
prominent, is properly called the stage of infantile dependence
because its outstanding feature is the dependence of the infant on its
caretakers. The only mode of object relationship possible in this phase
28 Chapter 2
is complete identification with the object of dependence. Fairbairn

termed the relationship between the infant and the mothering figure
"primary identification" in order to describe the process of incorporat
ing the object so completely that the distinction between ego and
object is blurred. This equation of incorporation with a type of identi
fication differs from customary psychoanalytic usage, in which iden
tification means becoming like an object rather than incorporating it,
as for example in Jacobson's theory (see chapter 1). Fairbairn called
this more common meaning "secondary identification" because it is
based on differentiation between ego and object, and he believed that
this type of identification becomes the prominent mode of object relat
ing in the "phase of mature dependence." In this stage objects are
sought that can be incorporated. Since the breast is the most easily and
usefully incorporated object, it is the preferred object in this phase.
Orality becomes salient because the mouth is the path of least resis
tance to the object. This fact makes orality a prominent feature of this
phase but, according to Fairbairn, does not justify postulating the pri
macy of orality over the nature of object relations. The child is oral
because it seeks the breast; it does not seek the breast because it is oral.
The infant's desire for the breast is inevitably frustrated when the
desired object does not meet the infant's needs. When the sought
object does not appear, the infant feels that its love, its sucking, has
destroyed the object. Once the infant makes such an interpretation, a
problem arises, for satisfaction appears to make the object disappear.
This interpretation leads to the fundamental conflict of this phase: the
infant seeks to incorporate the object but doing so "destroys" the
object. From the infant's viewpoint, its own desire for the object
threatens its existence. The result is a conflict between longing for the
object and fear of it and a tendency by the infant to withdraw from the
object in order to save it. Because of this conflict Fairbairn termed this
phase of his scheme the "schizoid position." He believed the schizoid
position to be the fundamental psychological position and considered
schizoid phenomena universal because frustration of some degree is
inevitable.
Fairbairn agreed with Abraham's division of the first developmen
tal phase into two subphases. Abraham (1924), who defined the stage
by its orality, subdivided it into an earlier, oral sucking, phase and a
later, oral biting, or sadistic, subphase. Fairbairn believed that the key
distinction in the two subphases is the emergence of ambivalence
toward the object in the second subphase as differentiated aggression
appears. Fairbairn called the first subphase "preambivalent" because
aggression does not yet appear as an affect distinct from libidinal
Fairbairn and Guntrip
longing. His conceptualization of the second subphase is another

instance in which his thinking was heavily influenced by Klein. As
will be seen in chapter 3, Klein (1937) believed that the second phase
of infancy is marked by the infant's awareness of its love and hate for
the same object, which she called the "depressive position." In
Fairbairn's view, it is when frustration eventually leads to differenti
ated aggression that the ambivalent subphase of infantile dependence
begins, as the infant is now aware of loving and hating the same
object. The danger to the object now comes from the infant's aggres
sion rather than its love, as in the preambivalent subphase. Like Klein,
Fairbairn conceptualized the depressive position as the intent to injure
the loved object. The problem of the schizoid position is to love with
out destroying; the problem of the depressive position is to hate the
loved object without destroying it.
In both the schizoid and depressive positions the object becomes
"bad," but the infant forms different interpretations of the "badness."
In the schizoid position, the object is loved, but when it frustrates, it
becomes a "deserter" and is experienced as "bad" by virtue of its
unattainability. In the depressive position, the object is also loved, but
when it becomes frustrating, it is hated. The infant desires to injure the
ambivalent object in the depressive position but has no such intent in
the schizoid position.
In both the schizoid and depressive positions, the infant is forced to
master the anxiety of object loss. In the schizoid position, the infant
fears that its love threatens the object. The only means the infant has
to manage the anxiety of potential object loss in this incorporative
("preambivalent") subphase of infantile dependence is to internalize
the object. For Fairbairn, the motive for all internalization is the effort
to control the bad object. In the depressive position, the internalized
object is ambivalently experienced, a condition that creates an intoler
able internal situation and leads to the need to "dichotomize" the
object. To protect the object, it is split into an "accepted" and a "reject
ed" object. "Rejecting" now becomes a crucial technique for mastering
the anxiety of object contact, and its use marks the beginning of the
transitional phase of development.
While Fairbairn believed that the orality observed in infants justi
fied its salient position in a conceptualization of the infantile depen
dence phase, he saw no such justification for the concept of anality. He
agreed that the "oral phase" has a libidinal object, the breast, but saw
no libidinal object of anality. The infant does not seek the feces. The
developmental period others refer to as the "anal phase" is character
ized by rejecting behaviors, but for Fairbairn anality was only a
30 Chapter 2
symbol. Fairbairn saw the child in this period as still dependent on the
object while using rejecting techniques to differentiate itself from it.
Because of this combination of dependence and rejection, Fairbairn
labeled this phase, which embraces the periods transitionally termed
"anal" and "phallic-genital," as "transitional."
The transitional phase, according to Fairbairn, begins with the use
of rejecting techniques to differentiate self from object. Continuing the
ambivalence of the second subphase of infantile dependence, the
infant needs both to accept and reject objects. If the object relation
ships of infantile dependence are satisfactory, the infant is able to
dichotomize the object, utilize rejection, and achieve differentiation.
This process has both internal and external aspects. Interpersonally,
the child now becomes capable of forming a relationship that is not
based on primary identification. However, the growing child is still
dependent on the mother, and it now confronts a new type of ambiva
lence. If the child is too close to the external object, it will be in danger
of regressively becoming identified with it, thus risking its newly
emerging differentiated sense of self; but if it moves too far away, it
faces the danger of abandonment. The conflict between engulfment
and isolation are the characteristic anxieties of the relationship with
the external object in the transitional phase. Fairbairn's view of the
relationship between the growing child and the caretaker in the
transitional phase is close to Mahler et al/s (1975) view of the separa
tion-individuation process (as discussed in chapter 1). Both saw the
developmental task of this phase to be the separation from the care
taker while maintaining a meaningful bond that is not threatened by
merger, or the obliteration of self-object boundaries. The dilemma for
the toddler in this phase, in addition to the isolation-engulfment con
flict, centers on the expulsion and retention of internal objects: expul
sion achieves autonomy but risks the emptiness of life without objects
whereas retention achieves fullness but risks loss of differentiation
from the object. According to Fairbairn, the child in this phase must
manage rejection and acceptance of both internal and external objects.
From a developmental perspective, the critical task of the transitional
phase is to reject objects without losing them; the child must learn to
form dependent object relationships while maintaining differentiation
between self and object. Successful completion of this task prepares
the child for the mutual dependency that Fairbairn believed was
characteristic of all mature object relationships.
Fairbairn gave little attention to the phase of mature dependence,
most likely because he postulated that all psychopathology originates
in the earlier phases. The importance of this phase, according to
Fairbairn, lies in its representation of the goal toward which all devel
opment is aimed. Successful resolution of the transitional phase
allows the child to maintain a dependent tie to a differentiated object.
The ability to form a nonincorporative object relationship requires
complete differentiation so that the object can be accepted for who
he/she is. Both the accepted and rejected objects must be "exterior
ized" for this to occur. When infantile dependence is completely relin
quished, the object can be given to and dependence can be mutual.
This mode of object relationship allows for genital sexuality.
Psychological Structure
According to Fairbairn, psychic division originates in the infant's

experience of the unsatisfactory object. Here again Fairbairn was
influenced by Klein (1952b), who emphasized ego splitting as a nor
mal process in early development to manage the anxiety of frustra
tion. According to Fairbairn, if the object provides a satisfactory
experience, ego integration and development are fostered, yet if an
infant were to experience complete satisfaction, no psychic division
would occur. All unsatisfactory experience leads to the internalization
of a bad object, which is both exciting and frustrating. Since the infant
seeks an object that is not responsive, it feels that its own love is unac
ceptable and an internalized rejecting object is set up within the psy
che. In an effort to diminish the intense pain and anxiety of feeling
its own love rejected, the infant splits the unsatisfying object into
"exciting" and "rejecting" objects, both of which are repressed.
Since for Fairbairn all objects give rise to a corresponding ego struc
ture, this internalization of objects results in a psychic division into
libidinal and anti-libidinal egos. The satisfying object, which he calls
the "accepted object," remains within the ego and becomes the "cen
tral ego." The "exciting object" becomes the libidinal ego, which, like
the central ego, is dynamic; however, the libidinal ego is relatively
more infantile and less reality oriented than the central ego. The
"rejecting object" becomes structured as the "internal saboteur" and
uses aggression produced from the frustration by the unsatisfactory
object to repress the libidinal, ego-exciting object structure. The central
ego rejects both the libidinal ego and internal saboteur, and in this
reaction it too uses aggression. The central ego deploys aggression to
reject the exciting and frustrating objects, or the libidinal ego and
internal saboteur, and the latter structure uses aggression to repress
the exciting object and its subsidiary, the libidinal ego. The configura
tion of the central ego, or accepted object, repressing the internal
32
saboteur-rejecting object and the libidinal ego-exciting object while the

internal saboteur represses the libidinal ego is what Fairbairn calls the
"basic endopsychic situation."
It should be noted that in Fairbairn's scheme the aggression
deployed for repression is wholly a result of frustration. Fairbairn saw
no evidence for an aggressive drive, or even aggression analogous to
libido, embedded in an ego structure seeking objects. His view was
that aggression appears only when libido is frustrated by an unsatisfy
ing object, resulting in ambivalence. To remove the danger to the good
object, aggression is deployed for the object splitting that is necessary
to manage the pain of frustration.
Fairbairn's (1954) case of Olivia illustrates his theory of the dynam
ics of psychic division. The patient was anorexic and agoraphobic. Her
developmental history revealed early feeding difficulties and constant
crying. Her father, unable to tolerate her crying, adopted the strategy
of holding her down until she stopped crying. This method worked so
well that Olivia learned to "hold herself down." Her father generally
interfered with her life by adopting an overprotective, stifling attitude
toward her. This interference excited her, but her father was also a
rejecting figure who inhibited her oral needs and general spontaneity.
Olivia "held herself down" by relentlessly attacking her own needs
from within. This internal rejecting object constituted a primary iden
tification with her father and was tantamount to her antilibidinal ego
attacking the libidinal ego that longed to have its needs met. The suc
cess of the antilibidinal ego resulted in the almost total inhibition of
Olivia's craving for the oral incorporative object, and anorexia was the
outcome of this massive repression. The exciting object, which was
also represented by the father, had to be split off from the rejecting
object and repressed owing to the intense frustration of needs.
This case illustrates Fairbairn's understanding of the formation of
psychic structure: it describes the internalization of the frustrating
object and its splitting into the rejecting and exciting objects, which
results in the division of the ego into the libidinal and antilibidinal
egos that split off from the central ego. Although Olivia was highly
symptomatic, Fairbairn believed that her dynamics differed from the
normal situation only in degree and that while people who are not
symptomatic do not repress the libidinal ego with such severity, the
repression, its motivation, and the resulting psychological division of
the ego are characteristic of the human personality.
Is Fairbairn's "basic endopsychic situation" simply a relabeling of
Freud's structural model of ego, superego, and id? Fairbairn believed
that his concept of mental structure was not a mere change of nomen
clature and that his model possessed distinct advantages over Freud's
concept of the tripartite division of the psyche. First, from a metapsy-
chological viewpoint, Fairbairn believed his theory to be more elegant.
He did not have to stipulate impersonal impulses and then explain
how ego structure could be derived from them. Fairbairn believed
that by divorcing energy from structure Freud required cumbersome
postulates of transformations of energy into structure that are difficult
to justify; in chapter 1, we saw how this problem informed the efforts
of the ego psychologists to account for the autonomy of the ego.
Second, Fairbairn regarded his theory as closer to the experience of
patients and criticized the "id" as an impersonal theoretical artifact,
not something actually experienced. Even more important for
Fairbairn (1943) was the fact that his theory, unlike Freud's, could
account for the attachment of libido to its object. To resolve the per
plexing problem of the stubborn attachment of libido to painful
objects Freud was compelled to adopt his highly speculative, ques
tionable concept of the death instinct, which ultimately explained lit
tle. With his object relations model Fairbairn pointed out that when an
object is repressed, an ego structure is formed and that giving up the
object is, consequently, losing a part of the ego, resulting in annihila
tion anxiety. By equating the internalization of objects with the forma
tion of ego structure, Fairbairn believed he accounted for the
adhesiveness of object attachments, including the clinically frequent
tenacity of ties to painful objects. This explanation has implications for
the concept of resistance, as will be discussed presently.
Fairbairn also believed that his theory employed a more satisfac
tory concept of aggression. According to Freud's dual drive theory,
repression uses aggression, but aggression is also a drive, concepts
which lead into the quagmire of trying to explain how one aggressive
drive can repress another aggressive drive. According to Fairbairn's
(1944) theory, aggression is a response to frustration and repression
becomes "one structure using aggression to repress another ego
structure with an aggressive charge" (p. 119).
Further, Fairbairn contended that his concept of the internal sabo
teur had a great advantage over Freud's concept of the superego. The
internal saboteur performs many of the functions of the superego but
in an amoral fashion. Because tempting, overstimulating objects excite
without satisfaction, they must be repressed. While Freud's superego
represses impulses that conflict with a moral sense, Fairbairn's inter
nal saboteur rejects threat and pain, and morality has no meaning to
it. According to Fairbairn, the superego is a higher level structure that
has little to do with the basic endopsychic situation—and even less
34
relevance for psychopathogenesis. As will be shown later, Fairbairn

believed that psychoanalytic theory made a fundamental error in its
emphasis on guilt in the theory of neurosis.
In accordance with his metapsychological reformulation and recast
ing of the stages of development, Fairbairn (1944) reconceptualized
the meaning of the Oedipus complex, a situation that he believed
arises as the child is able to relate to two objects rather than only one.
When the child's needs become genital, it experiences new frustra
tions because the genital needs are never satisfied. This ambivalence
results in the internalization of a bad maternal and a bad paternal gen
ital object, each of which is split into an exciting and rejecting object.
To simplify this complex emotional situation the child perceives one
parent as the exciting object and the other as the rejecting object,
perceptions that result in the Oedipus complex.
In Fairbairn's view, neither the desire for the oedipal object nor the
triangular situation produces guilt. "Pseudoguilt" issues from the
Oedipus complex to the degree that demand for parental love is not
fulfilled. From this rejection of its needs, the child concludes that its
own love is bad. Unfulfilled longing for love in the phase of infantile
dependence results in feelings of shame. If shame is the outgrowth of
that period, the unmet oedipal longings will also be experienced as
shameful, and this feeling will be masked by pseudoguilt. On the
other hand, if the object relationships from the infantile period were
satisfactory, there is little reason for guilt in response to unfulfilled
oedipal longings. This deemphasis on the Oedipus complex and guilt
in favor of shame and weakness in the dynamics of pathology is
another instance in which Fairbairn's views presaged Kohut's (1977)
later formulations of self psychology.
As might be guessed from this view of the Oedipus complex, Fair
bairn saw no major role for guilt in development. He believed that one
of the primary mistakes of the classical psychoanalytic theory of
development was an overemphasis on guilt. Hence, guilt is not a sig
nificant variable in pathogenesis. What, then, are the crucial variables
in the onset of psychopathology for Fairbairn? To answer this ques
tion, we now turn to his reconceptualization of the psychoanalytic
view of psychopathology.
Psychopathology
Fairbairn's (1941) theory of psychopathology was as developmentally

based as the traditional psychoanalytic view, but his reformulation of
crucial developmental issues resulted in a markedly different under
standing of every form of neurosis, character pathology, and psy

chosis. In accordance with his view of the schizoid position as the ear
liest developmental stage, Fairbairn (1940) believed that the schizoid
character is the most severe form of psychopathology. Since the
infant's earliest need is to have its love accepted, the most severe
trauma is the feeling that one's love is rejected, a feeling that arises
from excessively frustrating object relationships in the first subphase
of infantile dependence and that leads to feelings of shame, weakness,
and helplessness.
The lack of responsiveness from the object in the earliest develop
mental phase intensifies the normal schizoid position by producing
intensified desire to possess the object. If the deprivation is strong
enough, the desire becomes so strong and desperate that the infant
wishes to devour the object to secure it. To the infant, its the craving to
possess the object that now threatens to destroy it, a perception that
intensifies anxiety over losing the object. According to Fairbairn's con
ceptualization, this is the most painful anxiety a child or an adult can
face because it renders all object contact potentially destructive. Every
move toward the object elicits the fear of destroying it; the object can
be protected only by withdrawing from contact. The need to with
draw exacerbates the normal splitting of the ego as the libidinal ego
withdraws from contact with reality. In this emphasis on ego splitting
Fairbairn is again indebted to Klein. Although his concept of the schizoid
split has an emphasis different from Klein's concept of ego splitting, he
adopted Klein's (1957) view that ego splitting is characteristic of severe
pathology.
The withdrawal from all object contact is dangerous. As the emo
tional investment in object contact is withdrawn, the ego loses contact
with reality and begins to lose the sense of its own existence.
Clinically, this state appears when the patient talks of feeling nothing
or of feeling dead, as though he or she has ceased to exist. Need for
the object is intensified as a result of the deprivation, and the schizoid
can find satisfaction nowhere. Objects can be neither reached nor
avoided: Object contact imperils the existence of the object needed for
its survival; and withdrawal endangers the ego directly. Ultimately,
the schizoid fears the loss of the ego whether he or she moves toward
or away from objects. The dilemma of this dual anxiety results in the
feeling of futility so characteristic of the schizoid.
It was Fairbairn's contention that this schizoid dilemma—intense
need for the object and fear of destroying it through contact—results
in adult schizoid pathology. According to Fairbairn, the schizoid
patient possesses four basic attitudes. First, the schizoid is oriented
Chapter 2
toward partial objects, not whole objects. Owing to the fixation at the
oral level, others are treated like preambivalent breasts, as objects to
be incorporated for the patient's gratification. For example, when
Fairbairn (1940) asked a schizoid patient if he was happy in his mar
riage, "a look of surprise at my question spread over his face, followed
by a rather scornful smile. That's what I married for/ he replied in a
superior tone, as if that provided a sufficient answer" (p. 13). Fairbairn
believed that by adopting such an attitude, the patient was treating
others as he had been treated. Mothers of schizoid patients are either
indifferent or possessive toward their children, whom they do not see
as having value in their own right, and the patients take the easy route
of maintaining the relationship they know.
Second, such patients are oriented more toward taking than toward
giving, again as a result of fixation at the oral incorporative mode, but
also because they are so fearful of emotional contact that all emotion is
repressed and its contents overvalued. Eventually, all emotional
expression becomes associated with depletion, and giving is equated
with loss. Fairbairn mentions one patient for whom this dynamic
became so extreme that he was unable to pass an oral examination
because he was not able to give answers, even those he knew to be
correct. The schizoid can only take and must ultimately withdraw into
a remoteness detached from all emotional contact with others.
These two characteristics suggest how Fairbairn would view
patients who today would be characterized as having narcissistic per
sonality disorders. Clearly, Fairbairn viewed patients who treat others
as objects for their own gratification, who can only take but not give,
as schizoid. The self-absorbed and exploitive attitudes of the narcissis
tic orientation, according to Fairbairn's formulation, emanate from the
schizoid fixation at the preambivalent level of infantile dependence.
The patient can relate to objects only with an attitude of "primary
identification," an oral-incorporative attitude. To protect the object,
the patient withdraws from all object contact, but when he or she does
come into contact with objects, the ego attempts to incorporate them
in accordance with the only way it knows to form object relationships.
This dynamic is closely linked to the third schizoid characteristic,
which Fairbairn calls the "incorporative factor." Since schizoids
equate relations with others, or any form of giving, with depletion,
relationships with objects are transferred to the realm of inner reality,
where they have value. This preoccupation with depletion also char
acterizes the schizoid character's orientation to creative endeavor.
When such a person does create, he or she tends to regard the product
as worthless because to value the creation means to lose something
valuable and therefore to be impoverished. With this formulation,

Fairbairn offered an alternative viewpoint to a customary interpreta
tion. The devaluation of one's own products is typically viewed as a
narcissistic issue, reflecting self-devaluation or the protection of
grandiosity. For Fairbairn, it is most commonly the schizoid fear of
losing something valuable.
Among some schizoid patients, this issue is manifested by a "substi
tution of intellectual for emotional value" (Fairbairn, 1941, p. 20). Such
patients resort to an intellectual thought process in order to maintain
the overvalued emotions within, a process that results in an intellectual-
ized character style. Intellectualization, for Fairbairn, is not simply
a defense against affect; it is an effort to maintain a stranglehold on
overvalued emotions that the patient cannot bear to give up.
The fourth characteristic is the "emptying of the object." As a con
sequence of deprivation in the preambivalent object phase, the infant
feels it has emptied the breast by its incorporative strivings for it. To
contact the object is to empty it, thereby destroying it. Again, the
result is avoidance of object contact to preserve the object.
It is clear that in Fairbairn's discussion of schizoid characteristics he
reinterpreted many typical symptoms and character styles as manifes
tations of schizoid dynamics. One can see in his attitude toward nar
cissism and intellectualization, for example, a clear clinical alternative
to the traditional view of these symptoms. It was not simply a case of
his recasting impulses as object relationships; his view of psy
chopathology was founded on a basis different from all other psycho
analytic conceptualizations of pathogenesis. Specifically, Fairbairn
conceived of many pathological features, such as narcissism and intel
lectualization, as desperate efforts to avoid object contact. The
schizoid's presenting pathological picture hides the intense fear of
object contact that reflects the primitive incorporative mode of longing
for objects, a longing found so shameful by the patient. The clear clini
cal implication is that unless the clinician sees this schizoid root of the
presenting symptom picture, the depth and severity of the pathology
is likely to be missed.
Whereas schizoid pathology in Fairbairn's model is rooted in the
preambivalent phase of infantile dependence, depression is seen as
originating in the ambivalent subphase. If in the preambivalent phase
the infant's need to love is accepted and it feels its love is acceptable,
it will be able to feel aggression toward the loved object and in
the ambivalent phase will dichotomize the object into the hated
and loved objects. If however, the infant does not feel that its love has
been accepted by the object, in the ambivalent phase of infantile
Chapter 2
dependency it becomes anxious that its aggression will destroy the

loved object. Attempting to master the anxiety, it incorporates the
ambivalent object in a desperate effort to save it. Such an infant has
not mastered the ability to hate the loved object without feeling that it
is destroying it and is beset with unresolved ambivalent feelings
toward the incorporated object, which is now repressed.
The repressed ambivalent object determines further object relations
and is activated in the future by aggressive feelings toward the loved
object. Constantly fearful that its aggression will destroy the loved
object, the child suffers from chronic anxiety regarding object loss.
When later object relations evoke ambivalence, anxiety over loss of the
object may result in depression. Thus, in Fairbairn's view, the
repressed ambivalently valued object is the root of adult depression.
Nor is the depressive state considered a defense. The schizoid and the
depressed states are both primary psychopathological constellations
resulting from unsatisfactory object relations in the phase of infantile
dependence leading to the repression of bad objects. The schizoid
represses the exciting and rejecting object and so must avoid all object
contact. The depressive represses the ambivalent object and, unlike
the schizoid, can love and derive some degree of satisfaction from
object relations; however, relations with others will always be fragile
because aggressiveness can easily disrupt them.
All other forms of psychopathology, according to Fairbairn, are
efforts to manage the conflicts of the transitional phase rooted in the
schizoid or depressive positions. If there is unresolved conflict in
either the schizoid or depressive positions the transitional phase
becomes pathogenic. Both conditions involve fixation to objects in
the infantile dependence phase, which renders impossible the reso
lution of the transitional phase in the direction of differentiation and
the capacity to give to the object. Consequently, defensive tech
niques must be used in the latter phase, and these defenses consti
tute the various forms of neurosis. Fairbairn described four such
defensive techniques, each corresponding to a psychopathological
syndrome.
Since the developmental task of the transitional phase is to differen
tiate the object and surrender infantile dependence, rejective tech
niques are prominent. The incorporated object must be "expelled" to
achieve differentiation, but this expulsion results in separation anxiety
and a sense of isolation. Retaining the object, however, means primary
identification—and thus failure to differentiate from the object and
loss of any movement toward autonomy—and is experienced as feel
ing "shut in" or engulfed. According to Fairbairn, all phobic states are
defenses against one side of this conflict. Although he did not discuss
specific phobias in detail, it is clear that agoraphobia, acrophobia,
and phobias of darkness would be considered defenses against the
anxiety of differentiating from the object in the transitional phase.
Claustrophobia and relationship phobias would be defenses against
the anxiety of engulfment.
If there is intense anxiety over expelling the incorporated object, the
conflict can present itself as between emptying and retaining contents.
In this case, expelling is experienced as the anxiety of draining the
insides and retention as bursting from overflowing within. Mental
states, especially affects, are retained and overvalued to keep the
patient "filled up" and to thereby avoid the anxiety of loss. However,
this state of "fullness" produces the intense anxiety of bursting from
internal tension, which can become manifest in such symptoms as
somatic complaints. The patient oscillates between the push to expel
and the urge to retain, is paralyzed by the inability to make decisions,
and attributes exaggerated powers to thoughts. Whereas the phobic
externalizes the incorporated object and enacts the conflict with exter
nal objects, the obsessional keeps the object within and vacillates
between letting go of the object and retaining it.
Instead of treating both the accepted and rejected objects as either
internal or external, the patient may incorporate one object while
externalizing the other. If the accepted object is kept within and the
rejecting object is externalized, the latter becomes a persecutor and the
patient feels that the goodness within is under threat from without.
The anxiety of the inside being under attack from the outside is
Fairbairn's formulation of the paranoid state.
The reversal of the paranoid state would be internalization of the
rejected object and externalization of the accepted object. Such a state
results in an overvaluation of the external object and a clinging
demand for it to compensate for the rejected internal object. The result
is a feeling of badness within and goodness without. This configura
tion represents the hysterical state, according to Fairbairn. Because of
the historical importance of hysteria in the development of psycho
analysis and the considerable attention Fairbairn gave to the disorder
in his reconceptualization of psychoanalytic thought, his views on
hysteria are worth considering in detail.
The hysterical response to the dilemma of the transitional phase is
differentiated from the other neurotic techniques by the excessive
intensity of the exciting and rejecting objects, an intensity that results
in the repression of both (Fairbairn, 1954). The overexcitement and
overrejection of objects originates in the failure of the environment to
40 Chapter 2
meet the infant's needs to love and have its love accepted. In the
future hysteric, the trauma of this unfulfilled need results in the
enhanced desirability of the exciting object, which remains elusive,
and a commensurate need to reject the object out of the frustration of
craving an unresponsive exciting object. Aggression turns from the
object to libido and represses the libidinal ego with excessive severity,
creating a deep ego split, the symptomatic expression of which is the
dissociative state of the hysteric.
It is the mother—or, more accurately, her breast—that both excites
and rejects. The future hysteric reacts to this exciting and frustrating
object relation with a premature libidinalization of the genitals in
infantile masturbation. Fairbairn's view is that genitality enters the
picture as the overexcited child identifies the genitals with the breast.
To the hysteric, sexuality is oral, as can be seen in the clinging, incor
porative needs of the hysteric. Repression is not directed to genital
sexuality but to the objects of infantile dependence with which the
genitals are associated. When the child reaches the transitional phase,
it is still very much attached to these objects, a condition that compro
mises its ability to differentiate from them. In lieu of differentiation,
the child overvalues the external object, seeking the infantile object
under the guise of sexuality while repressing the bad object. Although
Fairbairn did not dispute the traditional psychoanalytic view that hys
teria involves repression of genital sexuality, he viewed the sexuality
of the hysteric as a fundamentally infantile dependent object relation.
In Fairbairn's view, oedipal conflicts assumed importance in the
clinical picture of the hysteric only because, in order to simplify a
complex emotional state, the child identifies one parent with the
accepting object and the other parent with the rejecting object.
Fairbairn considered oedipal conflicts to involve representations of
the traumatizing exciting and rejecting objects, repressed long before
the oedipal phase. Owing to the traumatizing early object relations,
the organ systems of the future hysteric become libidinally charged as
substitutes for the frustrating objects. When an external situation
breaks through the repression barrier, it revives the early trauma of
rejecting object relations and somatic symptoms appear as substitutes
for them. The hysterical focus on the body and bodily complaints con
stitutes, according to Fairbairn, a hopeless effort to achieve through
the body what had been frustrated in the early mother-child interac
tion. That is, the hysteric seeks an oral incorporative object relation
through the body.
Because of the premature erotization of the sexual organs, the hys
teric also seeks oral object contact through the genitals. Like somatiza
tion, oral eroticism, masturbation, and other autoerotic activities

result from object frustration. This premature genitalization of orally
sought objects creates the conditions for overvaluation of the exter
nal object in the transitional phase. The hysteric uses the body and
the overvaluation of the external object as the means for attempting
to achieve the object contact which was frustrated in the phase of
infantile dependence.
Fairbairn's (1954) formulation of hysteria is illustrated in his dis
cussion of his patient Olivia. As stated earlier, Olivia suffered from
anorexia and agoraphobia and eventually became withdrawn to
the point of almost total passivity. Her first years were character
ized by severe feeding difficulties and the father's holding her
down in response to her hunger cries. The mother tried breastfeed
ing to no avail and had difficulty first in getting Olivia to feed from
a bottle and then finding suitable food for her to take. Because the
mother was unable to feed Olivia properly, she became the first
object to be both exciting and rejecting to her, and when her father
held Olivia down, he also filled both roles. In addition, her father
teased her in a sexually provocative way, inhibited her freedom in
order to protect her sexually, and yet did not prevent her from
being sexually traumatized. Each parent assumed the role of an
exciting and rejecting object, but Fairbairn believed that the
m other's assumption of both roles in infancy represented the
"nucleus round which the hysterical personality is characteristically
built" (p. 113). The simultaneous excitement and rejection of the
infantile object accounts for both the orality of the hysteric and the
severe repression of the libidinal ego when the exciting object is
attacked mercilessly by the antilibidinal ego.
In her childhood Olivia's greatest delight was to be given the top of
her father's egg at breakfast. After her brother grew old enough to
compete for the egg, she abruptly stopped eating breakfast with her
father and assumed a distant attitude toward him. In analysis, Olivia
dreamed about eggs to represent her father's penis. The egg symbol
illustrates that "whereas the sexuality of the hysteric is at bottom
extremely oral, his (or her) basic orality is, so to speak, extremely geni
tal" (Fairbairn, 1954, p. 114). Fairbairn believed that the case of Olivia
illustrated not only the extreme repression of the libidinal ego in hys
terical states but also the premature excitement of genital sexuality,
thus leading to a fusion of orality and sexuality. This connection is
supported by another case of Fairbairn's involving a patient who ver
balized an experience of being "a baby at the breast" but also "wanted
something between my legs" (p. 114).
42 Chapter 2
In all four types of transitional-phase psychopathology, the patho

genic conflict involves overwhelming anxiety about object contact,
whether because love destroys or because aggression destroys in lov
ing. Except for hysteria, Fairbairn offered no explanation for why one
of these four defensive techniques is preferred in a given case. He did
make clear that the distinction between patients in the two primary
psychopathological conditions and patients who utilize one or more
of the four defensive techniques lies in the greater degree of trauma
suffered by the former group in the phase of infantile dependence.
Regardless of why a particular technique is preferred in the transi
tional phase, Fairbairn's contention was that the Freud-Abraham
equation of psychopathology with libidinal position missed the
crucial issue in each syndrome. Fairbairn concluded that, by miscon
struing defensive techniques as libidinal positions, the traditional psy
choanalytic view ignored the fundamental depression or schizoid
state beneath the defensive constellation. For Fairbairn, at the root of
all psychopathology is an unfulfilled childhood longing expressed by
an infantile ego, ashamed and anxious of its longings for dependence.
There is no higher level pathology produced by conflict between
impulse and guilt, as there was for Freud, who felt that only such con
ditions were analytically accessible. Such a fundamental reconceptual
ization of the psychoanalytic view of psychopathology has direct and
potentially far-reaching implications for treatment, and it is to this
issue that we now turn.
Treatment
Despite Fairbairn's extensive critique of psychoanalytic theory and his

radical transformation of it, he said little about the concrete therapeu
tic implications of his views until very late in his career. This is partic
ularly surprising in the light of his often-repeated statements that
impulse psychology limited and even distorted practice by encourag
ing an impersonal style of interpretation (Fairbairn, 1944). That there
was so little focus on clinical theory until later presents some difficulty
in interpreting Fairbairn's clinical views, but they can be discerned
from various comments throughout his writings and from his last
papers, in which he began to formalize a clinical theory (Fairbairn,
1958).
I have already indicated that Fairbairn's reconceptualization of all
psychopathological conditions seems to imply a technical approach
different from that suggested by drive theory. For example, the classi
cal analyst who views hysteria as primarily repression of genital sexu
ality due to oedipal conflicts will tend to interpret anxiety over sexual
ity as the basis for somatic complaints and hysterical emotional states.
Clearly, Fairbairn viewed such interventions as failing to reach the
depths of the patient's pathology. As we have seen, he viewed the
repression of sexuality in hysteria as a withdrawal of the libidinal ego
due to intense anxiety generated by object contact. Fairbairn felt that
the treatment of the patient must, at some point, reach this deep fear
of emotional contact in order to achieve a genuine therapeutic process.
Interpretation of the patient's repression of excitement as oedipal guilt
avoids this anxiety and thereby colludes with the patient's defenses.
Analysis informed by the impulse-conflict model tends to target
relief of guilt. Fairbairn's (1943) view was that hysterics, as well as
other patients, prefer to view underlying guilt as the source of their
problems because guilt defends against their fear of object contact.
To feel guilt over sexual longing is far more acceptable than to
acknowledge the fear that one's love is not good enough, along with
the resulting shame, withdrawal, and regressive longing for the
object of infantile dependence. The use of guilt to mask these painful
feelings is what Fairbairn called the "guilt defense," a defense that
tends to be fostered by the impulse-conflict model. One of the major
clinical implications of Fairbairn's theoretical views is the position
that guilt should be interpreted primarily as a defense against
schizoid withdrawal and regressive longing.
Since all intrapsychic conflict is between ego structures, according
to Fairbairn, the distinction between intrapsychic conflict and struc
tural defect is blurred. All people suffer some degree of ego splitting,
and the more severe psychopathological states differ only by their rel
atively deeper ego cleavage. Symptoms result from ego splits in all
patients, whatever the degree of pathology. Consequently, the fact
that structural defect is a component of psychopathology does not
imply a reduced accessibility to analysis. In the traditional viewpoint,
structural defects are not analyzable because only the id is dynamic,
the ego being conceived of as static. Fairbairn eliminated this problem
with his concept of dynamic structure, that portions of the ego are in
dynamic conflict with each other, each with a dynamic charge. Analy
sis of conflict between the libidinal and antilibidinal egos aims at reso
lution of conflict and concurrently heals an ego "d efect." For
Fairbairn, structure itself is dynamic; therefore, problems within it are
accessible to analysis. Criteria for analyzability cannot rest on the
assumption of an "intact ego," since no such ego exists.
As the splitting of the ego is always a component of the symptom
picture, the ego must always be addressed in clinical interventions
44 Chapter 2
(Fairbairn, 1944). It was Fairbairn's contention that interpretations

geared to impulses allow the ego to become a detached observer of the
clinical material. The patient can, and often does, assume the spectator
vantage point from which he or she may well elaborate the material
endlessly in order to defend against engagement. In this way,
Fairbairn believed, impulse psychology lends itself to collusion with
schizoid defenses. Fairbairn contrasted this type of intervention with
the more personal object relations interpretations that are addressed
to the experiencing ego. According to Fairbairn, this personalization
of interpretations allows less possibility for defense.
When Fairbairn (1958) finally did begin to formalize his theory
of technique shortly before the end of his life, it was this "personal
ization" of the psychoanalytic process that he most emphasized. He
ultimately decided to apply to the treatment process his theory that
ego development is based on satisfactory object relationships by
emphasizing the therapeutic importance of the personal relation
ship between analyst and patient. Indeed, Fairbairn went so far as
to state that impulse psychology cannot account for the analytic
emphasis on the transference. According to Fairbairn, if impulses
are of primary importance, their interpretation outside the transfer
ence should be sufficient. However, if symptoms are best resolved
within the patient-analyst relationship, a conclusion that is well
accepted by analysts, then problems with impulses are problems
within a personal relationship, and impulse psychology has no
tools for explaining this. However, from an object relations per
spective, emphasis on the transference is to be expected, as impulse
problems are symptoms of anxiety in object contact and can be
resolved only within an object relationship; it follows that in the
clin ical setting im pulse problem s can only be m eaningfully
addressed within the transference context.
This emphasis on the "personalization" of the analytic process led
Fairbairn (1958) to a reconceptualization of resistance. He pointed
out that analysis has changed emphasis from uncovering childhood
traumata to analyzing the current relationship with the analyst. To
Fairbairn, this change implied that the patient's resistance is not
directed against the unconscious past but against the inner reality of
the present. The patient defends against the repressed ego and
resists the analyst's efforts to bring this portion of the personality
into the treatment; the patient's resistance is directed against bring
ing his or her inner reality into contact with the analyst. That is,
what is ultimately resisted, according to this view, is not the process
but the analyst.
Similarly, Fairbairn conceptualized the transference as the patient's

effort to bring the analyst into his or her inner world of object rela
tions. The analyst attempts to alter the structure of this world, and the
patient not only resists the analyst as a threatening intrusion but also
perceives and experiences the analyst in accordance with this inner
world in order to maintain it. This view of transference and resistance
dictates a treatment focus on the transference as the here-and-now
representation of the patient's object relations inner world. In this
respect, Fairbairn anticipated crucial elements of Gill's (1981) view of
the analytic process (see chapter 8).
Although Fairbairn insisted on the crucial role of the transference in
treatment, he believed that the therapeutic action of psychoanalysis
lies in the whole of the patient-analyst relationship, not just in its
transference component (Fairbairn, 1958). In adopting this view, he
believed he was beginning to carry out the clinical implications of his
view that patients are crippled by the distorted object relations of their
inner reality which originates in the object relations of early child
hood. He reasoned that one cannot expect these inner relations to
change by interpretation alone, even if they are focused on the trans
ference. Interpretations can only be of benefit in the context of a rela
tionship with a "reliable and beneficent parental figure" (Fairbairn,
1958, p. 379). Such a figure allows for a satisfactory object relationship
that can begin to replace the early traumatic relationship. This empha
sis on the importance of the way the analyst relates to the patient is
similar to Winnicott's view of the analytic process (as will be seen in
chapter 4) and presaged Kohut's work on empathic attunement
(discussed in chapter 6).
This reasoning eventually led Fairbairn (1958) to yield some of the
traditional psychoanalytic strictures. He stopped using the couch
because he felt the couch technique interfered in the establishment of
a personal relationship between patient and analyst. He dismissed
arguments in favor of the couch as rationalizations designed to main
tain a distance perhaps desired by the analyst, but contraindicated for
the patient. Since the therapeutic efficacy of psychoanalysis, in his
view, is dependent on the maintenance of a positive object relation
ship between patient and analyst, it is the analyst's responsibility to
provide the setting most conducive to such a relationship. Fairbairn
opposed the restrictiveness of the traditional setting. He believed that
to maintain it was to put the purity of a prescribed method before the
well-being of the patient; that since the patient is the priority, restric
tions should be minimized and treatment method flexible enough to
adapt to the patient; and that to the extent restrictions are needed it
46 Chapter 2
should be acknowledged that they are for the analyst, not the patient.
Fairbairn believed that if a patient did not improve in a traditional set
ting, the analyst should ascribe the poor outcome to the failure of the
analysis to adapt to the patient rather than concluding that the patient
was unsuitable for analysis.
While the concept of a more flexible relationship may seem vague,
there are hints in Fairbairn's writings indicating what he meant. In
discussing a case of hysteria to explain his concept of conversion,
Fairbairn (1954) mentioned an incident in which the patient referred
to seeing a play the night before and he responded by commenting
that he had attended the program and had noticed that she was there.
Mention of this incident was dropped casually into Fairbairn's discus
sion of the case, but it indicates concretely by the very ease with which
he bent the classical framework, the flexibility in his technical
approach. According to the classical viewpoint, such a remark was a
violation of the analytic stance; from Fairbairn's vantage point, it was
a move toward the establishment of a personal relationship, without
which an analysis cannot succeed.
The good parental figure provides the environment for the release
and derepression of internalized bad objects without threat of destruc
tion. Since all internalized bad objects are ego structures, as they are
made conscious the ego begins to reintegrate. Thus, the personal rela
tionship between patient and analyst becomes a critical therapeutic
ingredient in ego integration.
As can be seen from this discussion, Fairbairn's (1958) approach
leads to a recasting of the aims of psychoanalytic treatment from mak
ing the unconscious conscious to repairing ego splits and reintegrating
the personality. Topographic shifts are meaningful, according to
Fairbairn, if they repair the ego but if unconscious impulses are ren
dered conscious without such a reintegration, they do not aid person
ality growth. Fairbairn recognized that some classical theorists, such
as Gitelson (1962), considered the goal of psychoanalytic treatment to
be the general maturation of the personality. While he welcomed such
discussion, Fairbairn's contention was that the drive-ego model was
unable to explain how psychoanalytic treatment causes general psy
chological maturation to occur. For Fairbairn, psychological matura
tion is tantamount to the healing of rifts in the ego, and therefore can
be explained only on the basis of an ego-object relations model of
personality change.
As discussed earlier, resistance, for Fairbairn, is the patient's con
tinued adherence to his or her inner reality despite the analyst's
efforts to bring it into the outer reality of the analytic relationship.
According to Fairbairn, the personal relationship is the only means the

analyst has to break through resistances. The transference is the
patient's "counter effort" to bring the analyst into his or her inner
world. As long as this inner world remains self-contained, it cannot be
changed, and the result is the "static internal situation" (Fairbairn,
1958). The patient maintains the internal configuration because the
patient is attached to the internal bad objects resulting from the frus
tration of early experiences that led to overwhelming anxiety and
hopelessness regarding contact with external objects and subsequently
to withdrawal to the inner world; this is the only object contact the
patient has and feels capable of having. The analyst attempts to breach
this closed system, but the patient attempts to hold to the transference
perception, that is, to the perception that is in accordance with his or
her inner object world. The analyst's leverage is derived from his or
her ability to form a personal relationship with the patient that does
not fit the transference perceptions.
The patient who is able to accept the analyst's "outer reality" has
moved from a closed to an open system. For Fairbairn, the transfer
ence, as well as the resistance with which it is identified, is resolved
more by the new relationship that develops during the analysis than
by insight. Nonetheless, analysis of the here-and-now transference is
at the center of Fairbairn's technical recommendations: transference
analysis is considered therapeutic when it leads to a new relation
ship that breaches the patient's previously closed system of object
relationships.
It is clear that Fairbairn had begun to develop a new model of psy
choanalytic therapy based on his object relations theory. This model
was in process when he ceased writing owing to his failing health.
Despite its incomplete nature, Fairbairn's clinical theory had already
provided new ways of understanding many types of psychopathology,
had eliminated the distinction between intrapsychic conflict and struc
tural defect and the use of this distinction to assess analyzability, and
had begun to revise the concepts of transference and resistance as well
as to reconceptualize the process and aim of psychoanalytic treatment.
Because Fairbairn's published work gives little indication of how he
put his theory into clinical practice, the most extensive case discussion
of his work is Harry Guntrip's (1975) description of his own analysis
with Fairbairn. His protege experienced Fairbairn as a detached "tech
nical interpreter" who was surprisingly orthodox and formal, show
ing little capacity for personal relatedness within sessions. Equally
surprisingly, Guntrip described Fairbairn as relying heavily on classi
cal theory, remarking to Guntrip at one point that the "oedipal
48 Chapter 2
complex is central for therapy, but not theory" (p. 451). Guntrip's dis
cussion of his analysis with Fairbairn seems to contradict the theoretical
model Fairbairn worked so hard to develop. Guntrip was well aware
of this, and he attributed it to both Fairbairn's personality and illness
(Fairbairn was sick during the major part of the analysis). Whatever
the cause of Fairbairn's behavior, his analysis of Guntrip raises the
question of whether he adopted his own model of treatm ent.
Especially troubling is Fairbairn's remark that the oedipal complex is
"central for therapy" inasmuch as he had based his theoretical and
clinical apostasy on his opposition to the centrality of this very con
cept. At minimum, we have far too few clues from Fairbairn as to the
clinical implementation of his model. For this, we must turn to his
most famous analysand, Harry Guntrip himself.
THE WORK OF HARRY GUNTRIP
Basic Concepts of Guntrip's Theory
Harry Guntrip, Fairbairn's protege and analysand, did more than any
other theoretician to popularize and extend Fairbairn's theory of object
relations. Viewing his own theory as an extension of Fairbairn's
thought, Guntrip (1961a) agreed with his mentor's contention that the
infant seeks objects, not pleasure, and that personality growth depends
primarily on the quality of object relationships. He also adopted
Fairbairn's view that the earliest developmental phase is infantile
dependence and that unsatisfactory object relationships in that phase
are the root of schizoid pathology, the most primitive psychological
state. Further, Guntrip concurred with Fairbairn's view that the ego is
born whole, if primitive, and that the result of unsatisfying object
relationships is ego splitting and the formation of psychological structure.
He agreed with Fairbairn's critique of an impersonal id, believing
that meaningful experience exists within the ego. Guntrip extended
Fairbairn's critique with his view that once Freud developed his con
cepts of narcissism, the distinction between ego and libidinal drives
was eliminated, rendering any distinction between ego and id mean
ingless. Further, the recognition of an unconscious ego in conflict with
the repressed unconscious indicated that split-off experience is part of
the ego. In Guntrip's (1971) view, the elimination of the id makes
ego psychology a personal, or human, psychology in which the ego,
interacting with others, is a personal center of experience. He con
trasted this concept of the ego with Hartmann's (1939) "system" ego,
an apparatus for controlling the drives.
Guntrip (1961a) was even more thorough and piercing in his attack
on the biological grounding of psychoanalysis than was Fairbairn. He
also pointed out that one does not see the operation of the pleasure
principle except in cases of severe pathology, when the personality
has broken down. Similarly, sexuality and aggression only become
problems within a fractionated ego owing to frustrated object rela
tions; when the ego is coherent and strong, sexuality and aggression
are experienced joyfully and help the ego grow. In adopting these
views, Guntrip anticipated many of the points Kohut (1977) would
later make in the development of his self psychology (as will be seen
in chapter 6, Kohut held similar views but used the concept of the self,
rather than the ego in his reformulation of psychoanalytic theory).
According to Guntrip, since the growth of the ego is dependent on
object relations, it is the relationship between ego and object that is cru
cial to the health of the personality, not the vicissitudes of the drives.
The natural progression of ego psychology, according to Guntrip, is
from a psychology of ego and id to one of ego and object.
Guntrip (1961a, 1971) was well aware that "object relations think
ing" could be construed as shifting psychoanalysis toward an inter
personal theory. He discussed the interpersonal theorists at length to
demonstrate that such a view was misguided. Focusing on Fromm,
Horney, Adler, and Sullivan, he pointed out that since these theo
rists viewed personality as a reflection of the social environment,
their theories are more sociological than psychoanalytic. Guntrip
attacked the "cultural pattern" theorists for abandoning the depth
psychology begun by Freud rather than building on it, and for ignor
ing the complexity of unconscious human motivation, especially in
psychopathology. This point is critical to Guntrip's clarification of
object relations theory and to the current resurgence of interest in
interpersonal theory. Since the publication of Greenberg and
Mitchell's (1983) Object Relations in Psychoanalytic Theory, the distinc
tion between object relations theories and interpersonal theory has
become blurred. Guntrip's critique of the "cultural pattern" theorists
makes clear that he would have been in clear disagreement with
such a merger of what he considered to be two different views of
human motivation. In chapter 7 the interpersonal theorists are dis
cussed in some detail, and the contrast between their views and the
ideas of the object relations theorists will be explained more clearly.
Suffice it to say that Guntrip believed interpersonal theory was
an abdication of psychoanalysis whereas what was needed was a
reconstruction of its theoretical edifice.
50 Chapter 2
Psychopathology
Guntrip took as his starting point for understanding psychopathology

Fairbairn's view that the infant's first need is to love and be loved and
that the first object relationship is organized around this need. If the
infant's need to love is rejected, it experiences the most painful emo
tional state: the feeling that its love is unacceptable. Like Fairbairn,
Guntrip (1969) believed that all psychological division originates with
the ego splitting that begins when the infant feels its love is rejected.
This unsatisfactory object experience leads to the internalization of the
bad object in an effort to master the experience. Guntrip, too, believed
the bad object was split into the exciting and rejecting objects.
However, he emphasized a different aspect of the bad exciting object:
he pointed out that the object that is longed for but not found is desir
able, but deserting, and he called it the "desirable deserter." This is
more than a change in nomenclature. Guntrip believed that the experi
ence of rejected love, whether in the adult or the infant, results in the
feeling that the object is deserting, rather than simply exciting, as
Fairbairn described it. This shift has significant implications for the
understanding of schizoid dynamics, as will be shown presently. For
now, the important point is that in Guntrip's schema the bad object is
split into the desirable deserting object and the rejecting object. The for
mer is tantalizing and, according to Guntrip, more painful than the lat
ter, because it continually frustrates and threatens abandonment
whereas the rejecting object is simply bad. The deserting object gener
ates the anxiety of object loss; the rejecting object, loss of the object's
love. Of the two, the ego much prefers the latter, as it is less threatening
to its existence.
Guntrip's writings provide illustrations of desirable deserters in
which the patient appears to be apathetic or depressed, whereas the
therapeutic material indicates intense, frustrated longing for objects,
often symbolized by dreams and fantasies of food. For example, a
school teacher who complained of depression and lack of interest in
school had a dream of going to a camp school where "the Head
walked away when I arrived and left me to fend for myself and there
was no meal ready for me" (Guntrip, 1969, p. 27). The patient went on
to describe his preoccupation with eating despite minimal food intake,
and then remarked that he typically eats alone and feels "totally cut
off" (p. 28). Guntrip interpreted the dream as reflecting both the
patient's longing for deserting objects, symbolized by the Headmaster,
and the resulting frustration, which leads to withdrawal that appears
to be depression. This example also indicates both the way signs of
apparent depression can be used to mask schizoid dynamics and the

frequently seen association between food and desirable deserters.
When the need to love is frustrated in the phase of infantile depen
dence, the infant fears loss of the object and seeks the object with
greater intensity. The more the need to love is frustrated, the greater
becomes the need to cling to the object. This is Guntrip's understand
ing of why patients cling to bad objects, a view that has, in fact, been
amply confirmed by the findings of ethological research (Bowlby,
1969). (More will be said about the connection between this line of
research and object relations theories in chapter 8.) According to Gun
trip's presentation of the dynamics of schizoid pathology, if love is
frustrated so that the anxiety of object loss becomes unbearable, the
infant longs to devour the object in order to keep it. However, this
desire leads to a more intense fear of object loss: the infant's desire to
devour the object triggers its worst fear—destruction of the object—
and now the infant itself is the potential source of the destruction.
Because the infant fears that its love will destroy the object it so des
perately needs for survival, it withdraws from object contact. This is
why, according to Guntrip, the schoolteacher's associations to his
dream, cited earlier, led to his feeling of being "totally cut off." The
patient so longed for the object represented by the Headmaster that he
wished to devour it, as symbolized by the preoccupation with gob
bling food; consequently, to preserve the objects of his desire he with
drew from all object contact. Guntrip's formulation of the schizoid
position is that "love gone hungry" leads to "love as destructive" and
the anxiety of destroying the object results in schizoid withdrawal to
preserve the needed object.
This pattern is shown clearly in the connection the schizoid patient
makes between need for the object and craving for food. Guntrip
(1969) discussed one patient who became ravenously hungry when
ever she saw her husband. She wanted to gobble food and drink in a
single gulp, yet in the presence of food she immediately lost her
appetite. More will be said later of the relationship between object
hunger and eating; the critical point for now is that hunger for the
object leads to the desire to devour it, to gobble it like food. "This anx
iety about destroying and losing the love-object through being so
devouringly hungry is terribly real. . . . The schizoid person is afraid
of wearing out, of draining, or exhausting and ultimately losing love-
objects" (Guntrip, 1969 p. 30). The woman who was hungry for her
husband commented that she had "an urge to hold him so tight that
he [couldn't] breathe, shut him off from everything but me" (Guntrip,
1969, pp. 29-30). She also expressed desire to kill the analyst.
Chapter 2
Because of the fear that desire for the object will destroy it, any
affective bond threatens the existence of the object. Because the ego
needs objects to survive, the anxiety of object loss is ultimately the
anxiety of ego loss. There is no psychological structure in this first
phase of development, so all anxiety is experienced as traumatic,
threatening the very existence of the ego. The result is a careful avoid
ance of all emotional contact with objects. However, this schizoid
withdrawal generates anxiety of loss of all contact with reality and
threatens, once again, the very existence of the ego. Objects are needed
as much as they are avoided. Consequently, object contact, though ter
rifying is sought. According to Guntrip, this situation leads to intense
"all or nothing" object relationships, which he called the "in and out
program." Schizoids can tolerate neither the presence nor the absence
of object contact. They desire to fuse with the object to secure it, but
this very desire threatens both the object and the ego and leads
ultimately to the futility so characteristic of schizoid psychopathology.
Guntrip (1969) formulated the schizoid conflict as the inability to
"be in a relationship with another person nor out of it, without in various
ways risking the loss of both his object and himself" (p. 36). Each side of the
dilemma risks ego loss, and this anxiety propels the patient to the
other side of the conflict. Guntrip illustrated the "in and out" program
with the following statement uttered by a nurse residing in a hostel:
The other night I decided I wanted to stay in the hostel and not go
home, then I felt the hostel was a prison and I went home. As soon as I
got there I realized I wanted to go out again. Yesterday I rang mother to
say I was coming home, and then immediately I feel exhausted and
rang her again to say I was too tired to come . . . . as soon as I'm with the
person I want, I feel they restrict me [p. 37].
It is common for clinicians to interpret such ambivalence as fear of

intimacy, and Guntrip would not disagree with such a description.
However, his contention was that such an extreme fear of intimacy is
a product of hunger for the object so powerful as to result in fear of all
object contact. Oedipal conflicts, however severe, cannot account for
the severity of this withdrawal.
Object contact is so threatening that a defensive character structure
is erected to protect the ego and allow an apparent affective bond with
a minimum of threat. According to Guntrip, there are two broad cate
gories of defenses that may be utilized in the service of defending
against schizoid withdrawal. The most common type of defense is
involvement in object fantasies that appear to be primary issues.
Guntrip regarded oedipal conflicts and longings for the breast as the
most typical of these issues. By including oral desires, Guntrip is
clearly extending Fairbairn's views another step. Both theorists saw
oedipal conflicts as primarily defenses against schizoid withdrawal,
but Guntrip believed longing for the breast typically served the same
function. For Fairbairn, longing for the breast is the underlying issue
for most schizoid pathology, but Guntrip believed such a longing
enables patients to defend against the underlying withdrawal from all
object ties by allowing them to believe that their wish is to possess the
breast. According to Guntrip, schizoid patients typically seek not the
breast but the cancellation of all object ties. The predominant under
lying fantasy is to return to the womb, to prenatal existence, when
nothing was longed for and nothing was needed (Guntrip, 1961b).
Guntrip did not make clear whether womb fantasies are a desire to
return to a "remembered" phase or are adult fantasies of what
intrauterine existence is like. In either case, it is fair to say that the
adult desire to achieve an objectless state does not imply memory of
prenatal life. Guntrip's point is that the adult schizoid longs to with
draw from all object ties and that his predominant fantasy therefore
becomes the return to the womb, the symbol of an objectless state.
To illustrate the way oedipal conflict defends against the regressive
longing to return to the womb, Guntrip (1969) described a patient
who suffered from "apparent depression," experiencing great diffi
culty finding interest in anything. Analysis uncovered a clear castra
tion fear that greatly exacerbated the apathy and resulted in crippling
difficulty performing daily tasks.
He lay in bed all day, curled up and covered over with bed clothes,
refusing food and conversation and requiring only to be left alone in
absolute peace. That night he dreamed that he went to a confinement
case and found the baby sitting on the edge of the vagina wondering
whether to come out or go back in, and he could not decide whether to
bring it out or put it back. He was experiencing the most deeply
regressed part of his personality where he felt and fantasied a return to
the womb, an escape from sheer fear of castration [p. 69].
Guntrip went on to state that the patient had, indeed, suffered

repeated castration threats from both his mother and an aunt that
were often accompanied with gestures with knives or scissors. The
"well-founded castration complex" was the result of his mother's hos
tility, according to Guntrip, leading to the patient's regressive with
drawal into himself. Although Guntrip acknowledged the castration
54 Chapter 2
complex, he believed that in itself it would not lead to such a crippling

illness. The desire to remain in a womblike state indicated that the pri
mary anxiety was not so much about losing the penis as about any
type of object contact. In Guntrip's view, maternal deprivation led the
patient to attempt to withdraw from all object contact, and the anxiety
about losing the penis was primarily a symptom of the patient's terror.
The other major category of defensive constellation against
schizoid anxiety reveals more clearly the underlying schizoid with
drawal. Some patients present with an aloof, superior, affectless,
apparently self-sufficient interpersonal style. The need to deny the
need for objects is so great that the defense involves a presentation of
complete indifference to others. The schizoid personality reveals in
this defensive constellation, neurotic defenses of oedipal and breast
longings, a lack of connection to objects. This defense is intended to
mask from patients themselves and from others their intense need for
dependence and the overwhelming anxiety associated with it. Object
contact is so threatening to the ego that it must believe in its total self-
sufficiency, resulting in a withdrawal into private grandiosity. This
defensive constellation is the essence of Guntrip's conceptualization of
what has over the past two decades come to be known as the narcis
sistic personality. (Guntrip's formulation of self-sufficiency as a
defense fits with Kernberg's concept of the narcissistic personality, to
be discussed in chapter 5, but Kernberg, 1975, believed the defense
was directed primarily against oral aggression rather than object
hunger. As will be shown in chapter 6, Kohut's, 1971, views differed
from both theorists inasmuch as he attributed stable grandiosity to
developmental arrest rather than to defensive need.) From Guntrip's
viewpoint, narcissistically aloof patients who believe themselves able
to do anything and who believe their need for others is nonexistent,
are defending against a terrifying longing to love and have their love
accepted. This desire in them has become so intense as to threaten
their very sense of self if any object contact is made. Consequently,
aloof withdrawal is the only way to maintain a sense of "autonomy."
The only alternative is merger which means loss of the sense of
personal existence.
Such patients typically display this defensive constellation in their
attitudes of self-sufficiency, superiority, and coldness. For example,
Guntrip (1969) discussed one patient who disdained children's games
and all other normal activities of youth. "As a child I would cry with
boredom at the silly games the children played. It got worse in my
teens, terrible boredom, futility, lack of interest. I would look at peo
ple and see them interested in things I thought silly. I felt I was differ
ent and had more brains" (p. 43). Such an attitude of superiority,
according to Guntrip, reflects the deep fear of object contact character
istic of the schizoid. Guntrip (1969) quoted another patient who
described his relationships this way: "I don't feel drawn to anyone. I
can feel cold about all the people who are near and dear to me. When
my wife and I were having sexual relations she would say 'Do you
love me?' I would answer: 'Of course I do, but sex isn't love, it's only
an experience.' I could never see why that upset her" (p. 44). These
attitudes of superiority and affectlessness are, Guntrip believed,
defenses against intense object hunger.
A personality structure of this type successfully defends against the
longed for merger between self and object. The defense allows the
patient to function and experience a minimal sense of connection with
reality without affective contact with objects, but requires constant
vigilance inasmuch as interactions with others continually threaten to
break through the defensive constellation by evoking affect. Since
interpersonal affect represents the most terrifying threat to such
people, their lives tend to be lonely, empty, and withdrawn.
Patients who are not able to defend so successfully become psy
chotic. In Guntrip's formulation of psychogenic psychosis the illness
represents the breakthrough of the desire to merge and involves a
blurring of self-object boundaries. The result is withdrawal from real
ity and concurrent self-world fusion. The psychotic is neither in the
world nor separate from it. By contrast, the schizoid personality, who
does successfully defend against the blurring of self-object bound
aries, maintains the boundary at the cost of withdrawal from the
world. Guntrip (1969) saw the schizoid character structure as a des
perate effort to "preserve the ego." From this point of view, all
nonpsychotic psychopathology, whatever the cost of the symptoms,
has the value of preserving the ego and, perforce, the connection with
reality.
It can easily be seen from this formulation that Guntrip viewed the
defensive structure of the schizoid personality as ultimately a defense
against psychosis. This is an important concept in Guntrip's theory; it
underscores the gravity of schizoid illness, as we shall see in the next
section, defines the goal of psychotherapy with such a patient as
reaching the core schizoid withdrawal underlying the defenses. If
treatment achieves its goal, it must run the danger of potential
psychosis.
Guntrip (1962) followed Fairbairn in the belief that when the infant
begins to feel ambivalence to the maternal object, the second phase of
infancy begins. For Guntrip, this is the phase in which the infant
56 Chapter 2
needs to know not so much that its love is acceptable but that it can
love without "destroying by hate." This sounds like Fairbairn's view
of two pathological positions, the schizoid and the depressed. In many
of his writings Guntrip appeared to adopt such a position, even at
times adding paranoia as a third psychopathological position.
However, Guntrip seemed ultimately to endorse no more than one
psychopathological position.
Guntrip (1962) agreed with Freud (1917) that depressives suffered
from repressed hatred of the loved object, but he questioned the tradi
tional psychoanalytic formulation that hate is rooted in the aggressive
drive (Freud, 1920). In Guntrip's view, because one can only hate the
object one loves and is attached to, hate implies an attachment, albeit a
frustrating one. Therefore, the opposite of love is not hate but indiffer-
rence; the object one hates is the object one seeks satisfaction from but
finds frustration in. From this Guntrip concluded that hate is frus
trated love, leading to the desire to attack and devour the object. To
preserve the object the hatred is repressed and turned against the self,
and depression results. The root of depression, then, according to this
view, is frustrated love, which leads to anxiety of object loss. But this
is essentially the same dynamic that Fairbairn and Guntrip proposed
as the root of schizoid pathology. This reasoning led Guntrip to the
conclusion that depression is a defense against schizoid pathology.
All his clinical vignettes and illustrations of depression were ulti
mately formulated as examples of the schizoid position as in the
clinical illustration described earlier.
Guntrip believed depression was the most significant defense
against schizoid pathology because of the strong preference patients
have for believing themselves to be depressed when they are, in fact,
emotionally withdrawn into a schizoid state. Indeed, Guntrip (1969)
broadened his discussion to a general human preference—to be found
in the history of ideas, including those of Freud himself—to view man
as bad, rather than weak. Guntrip's contention was that classical psy
choanalytic theory was dominated by the self-delusion that human
problems are due to guilt over "badness," a delusion that he believed
to be a denial of man's weakness. One of Fairbairn's most critical con
tributions to psychoanalytic theory, according to Guntrip, was the
movement away from this view and toward the understanding that
all psychopathology, whatever its presenting clinical picture, is a
product of fears with which the patient is unable to cope. Guntrip
pointed out that guilt implies strength but "badness"; that people pre
fer to feel they have the capacity to act, since this means they can con
trol their own fate. To admit that we would like to behave otherwise
but do not have the capacity is shameful because it exposes an inad

equacy: Freud's theoretical edifice, built on the guilty repression of
unacceptable affects, fits the human need to believe that one can and
colludes with the defensive needs of all patients who are fearful,
feel ashamed of inadequacy, and are unable to acknowledge their
incapacity.
Guntrip's view was that the rejection of love evokes feelings of
weakness because the infant longs for a response to the deepest need
it has but is unable to make the response occur. The infant's most pre
cious offering is inadequate, with the result that it feels weak, helpless,
and ashamed of its own needs. This "weakened ego" is fearful and
attempts to avoid exposure continually. To avoid recognizing this
painful weakness, the ego attacks itself for this very weakness, creat
ing the illusion that its badness is the source of its problems. Guilt is
superimposed upon the sense of badness and unworthiness, adding to
the feeling that the ego is undeserving and fostering the illusion that
the reason for unhappiness and symptoms lies in the feeling of unac
ceptability. Guntrip stressed that patients will quite readily admit to
feeling guilty because they are unconsciously relieved to avoid con
fronting their fears and weakness. In this sense, he believed that Freud
had not grasped the magnitude of his own discovery. While Freud
succeeded in demonstrating that mental life is filled with a vast array
of desires that it hides from itself, he saw the source of the uncon
scious as the unacceptability of these wishes. From Guntrip's point of
view, this conception does not go to the deepest root of man's lack of
self-awareness. Guilt masks the parts of the self the ego is most afraid
to know; namely, the feelings of weakness and helplessness and the
resultant sense of shame. All of this has profound ramifications
for treatment in general, as shall be seen shortly, but it has special
implications for the understanding of depression.
Recall that for Guntrip ego weakness always means the desire to
withdraw from all object contact and consequently evokes the threat
of ego loss. The patient turns his hatred against his own weakness in a
desperate effort to feel real. Guilt, self-hatred, and depression are the
price paid for this tenuous hold on reality. According to Guntrip,
depression is the prototype for all neuroses. All supposedly neurotic
conditions traditionally conceptualized as products of intrapsychic
conflict between affect and a moral standard are defenses against the
regressive longing to withdraw from all object contact and against the
resulting incipient loss of reality. Regarding the longing to return to
womblike security, Guntrip (1969) stated, "I regard this as the basis of
all schizoid characteristics, the deep secret flight from life, in seeking a
58 Chapter 2
defense against which the rest of the personality lands itself in a vari
ety of psychotic and psychoneurotic states, among which one of the
most important is depression" (p. 144). For Guntrip, the regressive
longing for the pre-object state is at the core of all depression.
The two patients cited earlier illustrate the regressive core of
depression. The schoolteacher who dreamed of the Headmaster walk
ing away and leaving him without food represents the schizoid with
drawal of an apparently depressed patient. The patient who had such
great difficulty getting out of bed demonstrates the regressive longing
even more clearly. Both patients suffered from depressive symptoms,
but in analysis the longing to return to prenatal security emerged.
Guntrip (1969) discussed in some detail a third case, a manic-depres
sive man who suffered prolonged periods of depression, characterized
by guilt and inactivity, alternating with phases of hyperactivity and
overwork. During the course of the analysis, the focus shifted from his
guilt over sex and aggression to his fears of facing the world. The
patient could not relax, had difficulty quieting his mind, and had
great difficulty sleeping as a result. He feared that if he "let go" in
sleep, he would never "get started" again. Guntrip used this as a clue
to the underlying "frightened self" seeking a retreat from life, and for
the first time the patient's schizoid characteristics were revealed. The
patient's "frightened child self" and a spate of fantasies and desires
to be warm and secure with the analyst emerged. He expressed a
longing to "let go," and during one session, he slept for about 40 min
utes. Guntrip's contention is that this type of material is typical of de
pressed patients if the analysis is allowed to proceed without
premature interpretations or interruptions.
Guntrip was able to fit addiction into his formulation that all psy
chopathology is rooted in the schizoid position. His view was that
when schizoid dynamics are operative, substitute objects are sought.
In some cases new human objects are used, but the despair in
schizoid withdrawal tends to draw the ego to nonhuman replace
ments such as food, drugs, and alcohol. Substances are more easily
controllable and are therefore unlikely to repeat the frustration
already experienced in human relationships. Further, it is in the
nature of schizoid withdrawal to prefer nonhuman modes of gratifi
cation. Recall that for Guntrip the goal of schizoid withdrawal is the
cancellation of all object relationships. The schizoid seeks nonhuman
forms of gratification in an effort to achieve the satisfaction missing
in early object relationships and in a manner that will not evoke
the anxiety and frustration of human contact. Food, drugs, and alco
hol all serve this purpose. Since the longing is for human contact,
however, the substance cannot succeed in providing the needed

gratification, and it must be continually sought in a desperate effort
to avoid the anxiety of schizoid withdrawal without the danger
of human contact. The result is substance addiction, which defends
against regressive schizoid longing by providing a sense of reality
contact. As soon as the substance wears off, the sense of reality threat
ens to disappear. Addiction, being a desperate effort to maintain a
sense of existence, becomes inevitable.
Eating disorders fit especially well into Guntrip's formulations.
Recall the connections alluded to earlier between the longing for the
tantalizing love object and food. As we have seen, the dreams of the
schizoid are frequently about food, and especially food that is not
eaten or eatable. The dream of the schoolteacher in which the
Headmaster left him without food is illustrative of this relationship.
Guntrip (1969) cites another patient who dreamed that she was enjoy
ing her favorite meal, only to have her mother snatch it away when
she came to the best part. The schizoid need to devour the object links
the "desirable deserter" with food and renders the latter a natural
substitute for the former, particularly given its relative availability.
According to Guntrip, in all eating disorders, food symbolizes the
tantalizing early object. It is a simple matter to extract from this con
ceptualization a formulation for each major form of food pathology.
Binge eating is an uncontrolled desire to secure the object by ingest
ing it. Bulimia has the same root but includes the fear of having
destroyed the object by devouring it, a fear that leads to the need to
expel the object to save it. If the craving to devour the object is over
whelmingly intense, it may threaten ego loss, resulting in an over
whelming fear of the longed-for object and a stubborn defense
against it. Anorexia is a schizoid defense against the desire to devour
the object. Typical of the latter is the patient, cited earlier, who
became hungry whenever her husband came home. She craved food,
desiring to devour it in a single gulp, but her appetite disappeared at
the sight of it. For such patients, the "in and out program" is enacted
in the domain of food and eating.
A similar assessment can be applied to the borderline personality
disorder. Guntrip did not use this label, but he described many
patients with the intense hostility, chaos, disorganization, overwhelm
ing anxiety, and intensely devaluing and idealizing object relation
ships so characteristic of borderline psychopathology. Without
detailed examination of the various symptoms of borderline psy
chopathology, Guntrip made clear that all these features were desper
ate efforts by the ego both to form such minimal object contact as the
60 Chapter 2
patient was capable of making, and to protect itself from being over
whelmed by any object contact. All the characterological features of
borderline pathology were seen by Guntrip as symptoms of either the
longing for merger or the fear of loss of the ego via withdrawal.
Patients in all these categories of psychopathology are beset by
internalized bad objects, the experience of which is usually extremely
painful. Nonetheless, psychoanalytic inquiry is not to be ended at the
uncovering of represented bad objects. Guntrip's crucial point is that
the bad objects are themselves defenses against ego loss. This may not
seem like a credible view, given the extreme pain bad object experi
ence can bring, especially in self-hating depressive and borderline
patients. However, Guntrip insisted that it is only the conception of
bad objects as defenses that can account for the stubbornness with
which patients cling to them. As painful as bad object experience is,
the pain provides a feeling of reality; the anxiety of loss of existence is
more threatening. The only alternative to bad object experience is
regression to a withdrawal so deep as to threaten the existence of the
ego.
In this view of psychological rock bottom, Guntrip parted company
with Fairbairn, who despite his emphasis on the schizoid position had
no concept of schizoid regression. It was his insistence on the impor
tance of longing for regression to the pre-object state that led Guntrip
to alter Fairbairn's theory of endopsychic structure.
Psychological Structure
Guntrip accepted Fairbairn's division of the psyche into the central,

libidinal, and antilibidinal egos; however, he questioned the concept
of the libidinal ego as the deepest, most repressed, ego structure,
being unable to accept the view that libidinal excitement by itself
could lead to severe repression and psychopathology. In a very real
sense, Guntrip came to believe that Fairbairn's view of the libidinal
ego as the seat of repressed material suffered from the same failing as
Freud's view of the id, namely that it could not account for severity of
symptoms; internalized bad objects; and, above all, psychological
weakness, shame, and feelings of inadequacy and helplessness, which
Guntrip believed were endemic to psychopathology.
As a result of the findings of his investigation into the depths of
psychopathology, Guntrip postulated a further division of the libidi
nal ego into an active, oral, sadomasochistic ego and a regressed ego.
The former corresponds, although perhaps imprecisely, to Fair
bairn's libidinal ego, and the latter is Guntrip's effort to explain the
schizoid core he saw in all psychopathology. The difference between

these two divisions of the libidinal ego is the distinction between an
ego invested in bad object relations and an ego seeking to return to
the prenatal safety of no object relations. The aspect of the libidinal
ego that is excited by objects is to some extent involved in a connec
tion with reality. Such an ego is possessed of breast and incest long
ings and fantasies. It has not lost its desire for objects; it seeks
primitive objects. This active ego has sexual and aggressive invest
ments in early objects, with all the conflicts and guilt resulting from
oedipal and preoedipal conflicts. However, the futility of unsatisfac
tory object relations, as discussed earlier, leads not to bad object
relations but to the longing to withdraw from all object contact.
As Guntrip (1969) states, this is a "passive regressed ego which seeks
to return to the antenatal state of absolute passive dependent
security" (p. 74).
The attack of the antilibidinal ego is directed against this passive
dependence, not against excitement. Guntrip viewed the self-hatred of
the depressive as a good illustration of such an attack: depressives
appear to be turning aggression from the object to the self, but they
are simultaneously attacking their longing for passive, absolute
dependence and defending against the awareness of it. According to
Guntrip, all people have this division of the libidinal ego and the
potential for withdrawal from object contact; those with psy
chopathology are characterized by a greater degree of dominance of
this regressive longing.
In Guntrip's view, the concept of the regressive ego is the final out
come of the shift in psychoanalytic theory from drive psychology to
object relations theory. Fairbairn had taken a major step in this direc
tion with his reconceptualization of psychic structure into ego divi
sions based on internalized objects. However, his concept of the
libidinal ego bases his theory of mental structure on a fear of excite
ment, rather that anxiety about object contact and, therefore, accord
ing to Guntrip, fails to address the depths of patients' anxieties.
Intense feelings of shame, weakness, and dependence imply a fear
that is far more powerful than excitement. In Guntrip's view, the con
cept of the libidinal ego as the deepest part of the psyche represents
Fairbairn's failure to move his theory far enough from Freud's drive
psychology. With his concept of the regressed ego, Guntrip believed
he had taken the final step in the development of the psychoanalytic
theory of psychopathology from a drive to an ego-object relations
psychology. It was this concept of the regressed ego that he felt made
sense of the fear and shame in the deepest levels of patients' psyches.
Chapter 2
Consequently, he believed this concept had far-reaching implications

for psychoanalytic therapy, and it is to the therapeutic process as
conceptualized by Guntrip that we now turn.
Treatment
Guntrip's clinical recommendations follow directly from his concept

of the regressed ego at the core of all psychopathology. On the basis of
his formulation that psychopathology is rooted in the desire to regress
to an egoless state, Guntrip believed that psychoanalytic therapy must
reach this regressive longing in the patient. In traditional analytic
therapy the aim of the process is to make conscious affects, impulses,
and wishes, and the mode of verbal interpretation is apposite to this
end. However, Guntrip, following Fairbairn's formulation of schizoid
dynamics, proposed a shift in the goal of the therapeutic process;
rather than bring discrete affects to consciousness, the psychoanalyst
should bring forth a split-off ego. This is not simply a matter of inter
preting correctly, but a process of allowing the patient to experience a
previously buried part of the self for the first time.
Once this component of the personality is apparent, the patient
must feel safe enough for its emergence in the analysis. The realization
of the existence of regressive longings is itself insufficient for thera
peutic efficacy; the buried personality must be experienced within the
context of the therapeutic relationship. (In this sense, Guntrip's
thought is in agreement with Weiss's, 1971, ego-psychological view
that unconscious material emerges when the patient feels safe. The
primary differences are that for Weiss unconscious material consists
of impulses and safety is equated with analytic neutrality, whereas for
Guntrip the unconscious is split-off portions of the ego, and safety can
be provided in other ways, such as by reassurance against abandon
ment.) With his prescription for the emergence of regressive longings,
Guntrip believed he was carrying out the direct therapeutic implica
tions of his and Fairbairn's object relations theory of psychopathology.
A consistent object relations approach to analytic therapy, according
to Guntrip, requires the relationship between patient and analyst to be
a critical component of the therapeutic process, as it was for Fairbairn
(1958). Guntrip went beyond Fairbairn by describing more specifically
the needed therapeutic relationship. Since the patient's problems are a
product of unfulfilled early object longings, the resolution must come
from a new object relationship that to some degree, satisfies the unful
filled childhood longings. In Guntrip's view, a part of the ego is "left
behind" and continues to cripple the growth of the personality. To
unblock the ego, its split-off portion must be reintegrated with the rest
of the personality; this requires a personal relationship that allows the
emergence and integration of the regressed part of the ego.
As we will see in chapter 4, Guntrip's concept of regression in the
analytic relationship is similar to Winnicott's notion that the analytic
process consists of belated ego maturation by the provision of a new
relationship. Winnicott, who was Guntrip's second analyst, had a pro
found influence on Guntrip. Although we will discuss Winnicott's
theory of technique in chapter 4, it should be noted here that when
Guntrip extended his theory of technique to the fostering of regression
to ego arrest, he blended Fairbairn's object relations theory with
Winnicott's technical emphasis on belated ego maturation.
According to Guntrip, the regressed ego is a product of trauma in
the child's longing for love and its offering of love. It is the original
unsatisfactory relationship that leads to the splitting off of the infantile
longings into the buried part of the self. If this traumatized component
of the personality is to reemerge, a new relationship must be experi
enced in which the child does not fear a reinjury to the infantile ego.
In this way the relationship the analyst offers to the patient becomes
crucial to the success of the treatment. The attitude of the analyst,
according to Guntrip, is critical for the provision of the safe conditions
necessary for the emergence of the regressed ego. The analyst must be
a demonstrably "better parent" than the original parenting figure.
In Guntrip's view, it is the analyst's ability to identify with the
patient that determines whether he or she can provide the atmos
phere necessary for the patient to take the risk of allowing the with
drawn component of the self to meet reality for the first time. Since
the patient must take a risk never before attempted, the analyst
must offer a relationship never before experienced. In emphasizing
the analyst's role in providing the necessary therapeutic relation
ship, Guntrip was again influenced by Winnicott, who (as we will
see in chapter 4), believed the analyst's role is to adapt to the
patient's needs. According to Guntrip (1969), what concerns the
patient is "whether the therapist as a real human being has a gen
uine capacity to value, care about, understand, see, and treat the
patient as a person in his own right" (p. 350). Only such a relation
ship has a chance of allowing the patient's long-buried individuality
to emerge.
Nonetheless, no sooner does the patient begun to experience this
emerging individuality than he or she experiences renewed threat.
The relationship offered by the analyst is itself a source of anxiety,
since the regressed ego now experiences the object contact it withdrew
64 Chapter 2
from the world to avoid. The threat of ego loss becomes very real, and
the previously withdrawn ego craves regression. As Guntrip (1969)
states, "The weak schizoid ego is in urgent need of a relationship, a therapeu
tic relationship capable of filling the gap left by inadequate mothering. Only
that can rescue the patient from succumbing to the terrors of ultimate isola
tion. Yet when it comes to it, the weak ego is afraid of the very relationship
that it needs"(p. 231). Guntrip points out that the ego cannot integrate
and grow without a relationship but that "the weakened ego always fears
it will be swamped by the other person in a relationship" (p. 231) Conse
quently, the patient is continually moving toward and away from the
analyst as he or she enacts the in-and-out program in the therapeutic
relationship. Each time contact is made with the regressed ego, it will
seek to rebury itself in the withdrawn state from which it is trying so
hard to emerge. This situation frequently leads to a therapeutic stale
mate. The therapeutic process then becomes a continual effort by both
patient and therapist to overcome this dilemma.
Two of Guntrip's clinical vignettes illustrate this enactment of the
in-and-out program in the therapeutic relationship. Guntrip (1969)
discusses one case in which the patient showed marked improve
ment after working through the trauma of having been rejected by
her grandfather, a trauma that had resulted in a fear of trusting oth
ers (p. 231). When Guntrip linked this fear to her relationship with
him, the patient initially showed further improvement. However,
one day, she arrived late, had not wished to come, and was hoping
to be able to leave before Guntrip could see her. When the interpre
tation was made that she was fleeing the therapist because she had
been more trusting of him recently, she responded this way:
She sat silent, pale, cold, uncommunicative, and then said that she had
told her mother that she felt her heart was a frozen lump inside her, and
she was frightened that she would never all her life be able to feel warm
and responsive and loving to any one. She added that she was afraid to
get too close to people, they were too much for her, and she revived a
fear . . . of becoming pregnant and having a baby, a fear which would
make marriage impossible [p. 232].
The patient went on to say, in response to Guntrip's interpretation

of her fear of closeness, that she was overwhelmed at the thought of
being close to anyone and could only feel safe at a distance. This
schizoid patient's flight reaction was part of her enactment of the in-
and-out program she had in response to a closer connection to her
therapist.
The second case illustration involves a woman who suffered from

a severe regression and occasional nightmares in which she would
scream for her mother and feel as though she were dying. As she
began to feel that Guntrip could fill the gap of utter isolation that led
her to seek her mother in her dreams, the nightmares diminished and
her condition improved. At this point two close friends of hers were
killed, and the patient began to feel that she was on the verge of col
lapse. She found it impossible to turn to Guntrip; she began to argue
stubbornly with him, and her nightmares resumed. However, she
then had a dream in which a headwaiter, who symbolized Guntrip,
was going away but then decided to stay on her account; in the same
dream she had a baby she had forgotten to feed. Guntrip pointed out
to her that she had been doing without him and had therefore been
repressing, not feeding, the baby in herself and that this withdrawal
had led to a fear of abandonment. Guntrip interpreted to her that
her increasing dependence on him led to anxiety over losing her
autonomy, thus motivating the push away from him.
This vignette demonstrates the typical schizoid response to thera
peutic improvement: the patient feels closer to the therapist, becomes
anxious, and then withdraws from the process. Both cases presented
here illustrate well the continual oscillation between connection with
the therapist and symptom improvement, on the one hand, and with
drawal and exacerbation of the clinical condition, on the other. Look
ing at the analytic process from this point of view, the concept of
resistance assumes a new meaning. Resistance in the process Guntrip
describes refers to the patient's continual effort to withdraw from con
tact with the analyst and return to the regressed objectless state. This is
Guntrip's application of Fairbairn's principle that the patient resists the
analyst not the analysis. Because the objectless state threatens such
sense of ego as the patient has been able to sustain, the resistance tends
to manifest itself most frequently in a "compromise relationship," in
which the patient can neither fully accept nor fully reject the analyst, as
to do either risks ego loss. The resistance may manifest itself in a vari
ety of defenses, such as intellectualization and isolation of affect, but
the value of the defense is always to maintain whatever sense of con
nection the patient is capable of. As was shown earlier, schizoid states
always serve a critical function for the patient: they maintain the reality
sense. In the treatment process the patient's defenses must be seen as
providing the only type of contact he or she can achieve, even as they
simultaneously block the progress of the treatment. In sum, the resis
tance must be analyzed, but it must also be appreciated as the form of
object contact the patient is currently capable of.
66 Chapter 2
Guntrip did exactly this in both of the aforementioned cases. To the

woman who pushed him away out of fear of betrayal, Guntrip inter
preted that she had come late and had sought to withdraw from him
in order to see if he could accept her independence rather than aban
don her in the face of her need for withdrawal. This is a good example
of what Guntrip meant by appreciating the patient's need for the
defense rather than simply interpreting it as a resistance to involve
ment with the therapist. Similarly, in the case of the woman whose
friends had died, Guntrip interpreted her need to argue with rather
than rely on him as her need for independence; he stated clearly that
he accepted this need and would not abandon her nor even change his
attitude toward her in response to it.
Despite the effectiveness of interpretations in these two cases, in
Guntrip's view interpretations themselves are insufficient for analytic
movement because making unconscious contents conscious does not
lead to the reintegration and growth of the ego. "Analysis must be
seen as 'exposing' a developmentally arrested psyche to the support
and new stimulus of an understanding relationship in which the ther
apist, like the parent, must wait while the child grows" (Guntrip, 1969,
pp. 178-179). This concept of resistance as a form of object contact and
of its treatment as the "exposure" to a new relationship presaged
Kohut's concept of allowing the patient's defenses to form the trans
ference relationship until they are gradually relinquished in the con
text of the new relationship (see chapter 6). Thus, Guntrip's concepts
of resistance and defense resolution anticipated some of Kohut's
fundamental concepts of the analytic process.
Both case illustrations also demonstrate Guntrip's view of the
transference. In the first case, the patient feared Guntrip would be like
her rejecting grandfather; in the second, the patient feared Guntrip
would not allow her independence without abandoning her, as her
mother had done. In both clinical situations the transference became
apparent as the patient began to feel the therapist was useful and
dependable. The anxiety of this potential view of the therapist was
stimulated in both cases by the image of the most painful, rejecting
figure of the patient's past. The figure most closely identified with the
patient's anxiety in object contact becomes the transference figure and,
as such, interferes with the advance of the treatment process. The
transference is the most stubborn obstacle to the progress of the rela
tionship between patient and therapist and, consequently, to the
appearance of the regressed ego. Its interpretation is crucial for the
new relationship between patient and therapist, which is necessary for
the development of the therapeutic regression. For example, Guntrip's
interpretation to the first patient that she feared that he would be like
her grandfather allowed her to begin once again to depend on him.
These case examples also show the ease with which Guntrip inter
vened noninterpretively. In the first case, after interpreting to the
patient that she feared he could not allow her autonomy without los
ing touch with her, he told her that he would be there when she
needed him. In the second case, he was even more direct and reassur
ing: after interpreting that the patient feared she would lose her inde
pendence, Guntrip (1969) told her that her dependence on him "did
not aim at robbing her of independence based on inner strength, and
[that he] could accept her independence as well as her dependence"
(p. 235). Guntrip believed such reassuring remarks helped foster the
patient's feeling of safety and security in the patient-therapist rela
tionship that is required for the eventual emergence of the therapeutic
regression. These examples demonstrate that Guntrip's concept of the
therapist's role is a plain departure from the traditional near-exclusive
reliance on interpretation. Guntrip clearly believed that the priority
for the therapist is to make contact with the regressed ego rather
than offer the most apt interpretation and that any intervention that
furthers that goal is preferable to interpretation.
Guntrip refers to three stages in the analytic process. Initially, the
oedipal object-related conflicts, involving sexual and aggressive feel
ings and guilt are addressed. However, according to Guntrip, such
dynamics are always to some degree defenses against the exposure of
the weak, immature component of the personality with its infantile
dependency. Consequently, when the oedipal conflicts are inter
preted, the infantile dependency longings and the shame associated
with the regressed ego threaten to appear and the schizoid compro
mise becomes manifest. The treatment task at that point is to over
come this stalemate. This critical breakthrough can be achieved only
by a combination of the safe therapeutic atmosphere, and the thera
pist's persistent interpretation and appreciation of the patient's resis
tance to the emergence of the regressed ego. The third phase of
treatment involves regression to the infantile ego and the rebirth of
the ego as a reunited whole. The therapeutic action, for Guntrip, lies
in regression and rebirth rather than in the content of interpretations.
In this sense, Guntrip reconceptualized the therapeutic process of psy
choanalytic treatment, and in so doing he believed he had arrived at
the logical clinical application of Fairbairn's and his own object rela
tions theory of mental structure and psychopathology.
Guntrip was realistic in his assessment of the possibility of such
a radical treatment in every case of psychoanalytic therapy. He
68 Chapter 2
acknowledged that such a deep treatment was often not possible, but
he believed that his concept of the therapeutic process provides the
therapist with a model that fits the patient's needs and issues far bet
ter than does the traditional interpretive model. Even though every
analysis will not be carried through to the deepest point, the model
Guntrip proposed sensitizes the therapist to the defensive nature of
such seemingly bedrock issues as oedipal conflicts and internalized
bad objects. The realization that even such painful issues can be used
as defenses against a still deeper regression, with longings to devour
and return to a womblike existence, helps the analyst guide the ther
apy to a deeper level, even though a complete regression and rebirth
may not occur.
Perhaps the clearest view of analytic therapy conducted according
to this model was presented by Guntrip in his lengthy discussion of
the manic-depressive case described earlier. The patient led a miser
able life, oscillating between the sluggish inactivity of depression, at
times accompanied by guilt, and hectic overwork, punctuated by sex
ual and aggressive outbursts. In the treatment process it became
apparent that the guilt was a feeling of contempt directed at his weak
ness and inability to function. He used somatic symptoms to defend
against the depression, often falling ill for long periods during which
he could do almost nothing. Guntrip used the somatization to focus
the treatment effort away from his guilt over his "bad" sexual and
aggressive impulses and toward his feeling of weakness and his wish
to withdraw. The patient acknowledged that he had to force himself
to function against deep feelings of inadequacy and fear. It will be
recalled that he was unable to relax for fear that if he "let go" he
would lapse into a completely nonfunctional state. During therapy the
patient came to understand that his manic drive was a desperate
attempt to avoid such an outcome; ultimately, he feared the recogni
tion of his infantile, regressed ego and used the mania to maintain
contact with reality. The crucial intervention at this point was
Guntrip's (1969) interpretation of the patient's apparent depressive
guilt as a component of his "self-forcing" and therefore, as a defense
against the recognition of his frightened child' self who was in a state of
constant retreat from life" (p. 158). After three and one-half years of
analysis, the patient entered a period during which his schizoid char
acter came to the fore; then after a retreat to the most regressed part of
himself, he embarked upon a course of steady and irreversible
improvement. The regression began when the patient evinced no dis
cernible reaction to his father's death, remarking that his father cared
little for children and adding that he himself was quite sensitive to
their sufferings. Guntrip remarked that the patient inflicted great suf
fering on "the child within." The patient replied that he had the wish
that Guntrip would put him to sleep so that he could awake and find
all his troubles solved. This comment, which marked the start of the
patient's recognition of his desire for regression and dependence on
his analyst, led to their intensification. At this point, the patient began
talking frenetically and dreamed of a man buried alive in a coffin.
Guntrip pointed out that the withdrawal he feared was so deep that
he felt as though he would be buried alive and his rapid speech
defended against this recognition.
The transference became the focus of the patient's schizoid conflict
between yielding to the longing for infantile dependence and forcing
himself into pseudoadulthood. He had fantasies both of smashing
Guntrip's car and of being in the car with Guntrip, of leaning against
him and putting his arm around him or of curling up in the back seat
as Guntrip drove. Guntrip (1969) interpreted these fantasies as fol
lows: "His hostile resistances to me and fantasied aggressions against
me earlier on were clearly a defense against his fear of helpless
dependence on me, and masked a fantasy of a return to the womb"
(p. 159). In a crucial session the patient began with an attitude of
impersonal, unfeeling detachment, and Guntrip pointed out that this
was a withdrawal in fear from his deepening attachment, depen
dence, and trust of him. The next session began the patient's steady
improvement; he reported a dream of going through a tunnel, which
Guntrip interpreted as a fantasied return to the womb. In the same
session, the patient expressed a desire to fight, because "a love rela
tionship is smothering " (p. 160). Guntrip interpreted this wish to
do battle with him as the patient's defense against his developing
feelings of closeness and dependence.
Subsequent material continued the theme of the patient's wish to
surrender to Guntrip's care (for example, he reported feeling greater
comfort on the couch) and his desire to flee in fear. However this tol
erance for his dependence and regressive longings irreversibly
increased, as did his clinical improvement. In one session he slept
most of the time; in the next analytic hour he reported that the previ
ous session had changed him by giving him a feeling of greater calm
and strength. Guntrip reports that the manic-depressive condition
never returned and that eighteen months after the case report was
written the patient's gains were still evident.
This case, which he reported in great detail, provides a good illus
tration of how Guntrip applied his treatment principles. One can
see that Guntrip interpreted the patient's guilt as a defense against
70 Chapter 2
regression. By remarking on the patient's abuse of his "child within,"

Guntrip elicited his first acknowledgment of his dependent and
regressive longings. After that point the treatment took on the charac
ter of the in-and-out program, with the patient seeking both complete
dependence and total detachment. Womb fantasies appeared as a
manifestation of the regressed ego, and eventually the regression
achieved its deepest enactment when the patient slept for most of a
session. The transference manifested itself most importantly as the
patient's deep need for and fear of trusting and surrendering himself
to the analyst, and its interpretation served to help the patient over
come this fear. When the patient trusted the analyst enough to allow
the regressed portion of the ego to appear in the analytic relationship,
therapeutic movement occurred.
In this case and others like it, Guntrip brought to fruition his and
Fairbairn's far-reaching theoretical revision of Freud's concept of
human motivation, development, and psychopathology. Guntrip's
description of the therapeutic process indicates that he, unlike
Fairbairn, viewed object relations theory as providing the rationale for
a clearly different, but nonetheless psychoanalytic, model of psycho
dynamic treatment, a model he implemented in his work with
patients. Guntrip went beyond Fairbairn in developing this model in
some detail. Guntrip believed that in his application of his and
Fairbairn's reconceptualization of psychoanalytic theory, he was able
to demonstrate how psychoanalytic therapy is conducted on a human
object relations, rather than a drive, model.
CRITIQUE
Fairbairn developed the most systematic object relations theory of

personality and pathology. He was the first theorist to build a the
ory on the concept of autonomy of object attachment, and his view
that object seeking is a primary function has been supported by the
research on attachment in animals and children, as we have seen in
chapter 1 (Bowlby, 1969). This object relations theory is closer to
patients' experience by avoiding the reductionism of classical the
ory. This type of theorizing was later lauded as "experience-near"
(discussed in chapter 6) by Kohut (1977). Fairbairn must be given
due credit for the development of a comprehensive theory of per
sonality development and pathology containing a minimum of
unfulfilled nonexperiential concepts. His "personalization" of psy
choanalytic concepts is a forerunner of recent efforts to make psy
choanalytic theory more of a psychology of persons rather than of
impersonal forces (for example, Mitchell, 1988; Stolorow, 1985).
Another of Fairbairn's major contributions is his emphasis on the

dependency-autonomy conflict in the dynamics of psychopathology.
Fairbairn was a pioneer in the recognition that the failure to resolve
the need to both attach and maintain independence accounts for many
human problems. The influence of Mahler and her colleagues has led
to wide acceptance of separation-individuation as a key developmen
tal process and force in pathology, but Fairbairn understood the
significance of this developmental trajectory before Mahler's work
appeared. Fairbairn also used it to understand neurotic pathology,
whereas Mahler and her coworkers have confined their clinical
application of the concept primarily to severe pathology.
Perhaps the most significant of Fairbairn's and Guntrip's contribu
tions to the theory of psychopathology is their appreciation of
schizoid dynamics in a range of pathological conditions, including
apparent neurosis. Both Fairbairn and Guntrip pointed out that the
overuse of guilt and structural conflict in psychoanalytic formulations
often masks a deeper level of pathology characterized by shame, fear,
and feelings of inadequacy, and both theorists were instrumental in
applying these concepts to a variety of symptomatologies. They
pointed out that treatment must often touch a deep level of schizoid
withdrawal, fear of object contact, and shame in order to reach the
patient. This view has been adopted to some degree by several more
recent psychoanalytic movements, such as self psychology (Kohut,
1977; Bacal and Newman, 1990), Winnicott and his followers (Khan,
1974), and interpersonal theory (Mitchell, 1988). Fairbairn and Guntrip
must be considered leaders in this general thrust of psychoanalytic
theory toward the appreciation of psychological deficit. In addition,
they developed a unique approach to deficit with their theory of
schizoid dynamics.
As is frequently true of theories, the strength of schizoid dynamics
is also its most glaring drawback. In their zeal to underscore its impor
tance, Fairbairn and Guntrip made the error of reducing all pathology
to the schizoid position. While Fairbairn theoretically recognized two
pathological positions, the schizoid and the depressive, his discussion
of pathology (and especially his limited case material), focused exclu
sively on the schizoid fear of destroying the object with love. Guntrip
yielded all pretense of belief in any other dynamic by formulating
depression as a schizoid condition. Ultimately, Fairbairn and Guntrip
reduced all human problems to a single issue. The drug addict, the
demanding borderline patient, the anxiety neurotic, and the depres
sive all suffer from the same "disease"—fear of object love. The the
ories of Fairbairn and Guntrip cannot provide an explanation for why
these syndromes are different, and they are limited in their ability to
72 Chapter 2
explain the particular issues in each clinical condition and the individ
ual dynamics of each patient. Their homogenization of pathology is
undoubtedly a primary reason for the limited influence of their theo
ries. Although recent interest in object relations thinking has led to a
greater degree of recognition for their work, their impact has not been
extensive even among relational theorists with few exceptions (e.g.,
Mitchell, 1988, and Celani, 1993).
Fairbairn's work is so theoretically dominated that there is insuffi
cient clinical material to see how his theory works in practice. While
the clinical work Guntrip reported is often impressive, it suffers from
the same limitation as Fairbairn's object relations theory: its applica
tion of one basic principle. For Guntrip, the analytic task is to meet the
needs of the patient's regressed ego, and while he reported some cases
in which this technical approach seemed to be quite successful, it is
the only type of intervention he advocated. One is left with the
impression that Guntrip believed in meeting a need for regression in
all patients, an error of homogenization of pathology in the clinical
arena. If Guntrip believed that some patients could be treated in
another way, this possibility is not indicated in his writing.
Fairbairn and Guntrip are primary advocates of what Mitchell
(1988) has called "developmental arrest theory," the belief that pathol
ogy is a block in the developmental process that must be undone in
treatment (see chapter 7). Unfortunately, their approach to the concept
of developmental arrest is so narrow in its single-issue focus on
schizoid withdrawal that the range of applicability of their theories is
limited. It remained for Winnicott to create a developmental arrest
theory which took into account a variety of developmental issues that
could be the source of a pathological outcome. Winnicott's views have
a strong connection to the theories of Fairbairn and Guntrip, but
they will be postponed until chapter 4, after the presentation of the
theoretical system of Melanie Klein, by whom he was also greatly
influenced.
CHAPTER 3
The Work of Melanie Klein
M e l a n i e K le in d e v e lo p e d t h e f i r s t s y s t e m a t i c b r a n c h o f p s y c h o a n a -
lytic thought that retained Freud's concept of the dynamic uncon
scious and remained within the psychoanalytic movement. Her ideas
are difficult to categorize because she based her thinking on innate
drives but always emphasized the significance of object relationships
in development and psychopathology. She stressed the importance of
the ego in development but advocated an aggressive clinical style of
interpreting impulses rather than systematically interpreting defenses.
Despite the confusion surrounding her work, she is considered here—
as she often is (for example, Greenberg and Mitchell, 1983, and
Guntrip, 1971)—as an object relations theorist because she viewed
development, despite her emphasis on drives, as organized around
the vicissitudes of object relationships and conceptualized the dynam
ics of all forms of psychopathology as object relationship conflicts. She
differed from most other object relations theorists, such as Fairbairn
and Guntrip, in the importance she gave to drives in the development
of object relations and in her aggressive interpretive technique.
Unlike many other psychoanalytic theorists, Klein developed a sys
tematic conceptual framework that she used to account for all clinical
syndromes and symptoms. The basis for her theory is Freud's (1920)
dual-drive formulation, as set forth in Beyond the Pleasure Principle.
Klein (1946, 1958) adopted Freud's view that the infant is born with
both libidinal and destructive impulses and that the ultimate fate of
the personality rests with the development of these drives and the
relationship between them. Indeed, Klein believed that analytic theory
had erred by focusing too much on libidinal drives and failing to
grasp the critical importance of the aggressive drive. She believed that
the aggressive drive is more potentially pathogenic than libido and
that psychoanalytic intervention at the deepest layers of mental func
tioning cannot be conducted without focusing on aggressiveness. To
gain a clear view of Klein's understanding of psychopathology, one
must approach her theory from a developmental perspective.
73
74 Chapter 3
DEVELOPMENT AND PSYCHOPATHOLOGY
The Paranoid Position
According to Klein (1935, 1948b), the problem in the earliest phase of

development stems from the fact that the infant is born with the
aggressive drive, giving rise immediately to annihilation anxiety.
Aggressiveness, for Klein, is the death instinct, or the drive to destruc
tion; thus, the ego is born with the anxiety of its own destructiveness.
Although the infant is also born with the life instinct, or libido, this
positive force is not inherently strong enough to dissipate the death
instinct completely; therefore, the primitive ego must use the mecha
nisms at its disposal to assuage annihilation anxiety. The first and
most dependable mechanism is projection. The infant attributes its
own destructiveness and the attendant anxiety to the breast, which
frees the primitive ego from the anxiety of being destroyed from
within (Klein, 1948a, 1957,1958).
Klein (1957,1958) drew two key implications from this view of the
origins of psychological life. First, she pointed out that innate aggres
siveness leads immediately to an object relationship. Because of the
immediacy of the projection of destructive impulses, mental process
from its origin has an object. Consequently, Klein disputed Freud's
notion of the stages of autoerotism and primary narcissism. While she
did not disagree with the notion of objectless states, she viewed them
as coexisting with the early relationship with the breast, and thus
rejected the idea of objectless stages of development. Second, she
believed that the cost to the primitive ego of the projection of aggres
siveness onto the breast is the preconception of a new state of danger
from without, as the breast holds the threat of destruction that once
resided within the ego. All the vicissitudes, mechanisms, anxieties,
and feelings associated with this earliest object relationship of a primi
tive ego fantasizing attack from a bad breast constitutes the "paranoid
position," the earliest developmental phase (Klein, 1948a, 1957).
Because of the immediacy and importance of the ego's early rela
tionship to the object, Klein (1935) concluded that the infant's mental
development from the earliest phase is a function not only of libidinal
position but also of the nature of the object relationship: "For where
we deal with etiology it seems essential to regard the libido-disposi-
tion not merely as such, but also to consider it in connection with the
subject's earliest relations to his internalized and external objects.. .."
(p. 267). For Klein, a good object relationship is a bond in which the
infant has an overall feeling of contentment and satisfaction. Only one
Melanie Klein 75
component of this object relationship consists of feeding. The defenses

are also crucial to the internalized object relationships that form the
ego and superego even in this early phase. Furthermore, Klein (1952a)
noted that infants as early as the second month will interrupt feeding
to look lovingly at the mother. From such infant behavior she con
cluded that the infant derives as much gratification from the "object
that gives the food as the food itself" (p. 96). Indeed, Klein (1952a)
believed that direct observation of infants supported the view that a
good object relationship makes feeding more gratifying. She pointed
to the fact that "sleepy satisfied" babies suck better if the early object
relationship is positive and that the "good suckers" become less
aggressive and greedy if the initial object relationship is good.
In this context Klein (1958) developed her position on the relative
importance of constitution and environment. She felt that the emo
tional well-being of the infant is dependent on its ability to form posi
tive early object attachments and that this, in turn, depends on the
relative balance of destructive and libidinal impulses. While frustra
tion is a major contributor to the amount of aggression in the early
object attachment, the infant's ability to tolerate frustration is the other
key variable. Both Zetzel (1951) and Guntrip (1961a) criticized Klein
for neglecting the role of the environment. Mitchell and Greenberg
(1983) adopt the view that Klein saw aggressiveness as originating
within and libido as originating without. However, Klein (1937,1952a,
1957) made clear that she believed that both libido and aggression are
constitutional drives greatly influenced by environmental factors
and cannot be separated from the vicissitudes of object relations.
Development within the paranoid position is a function of the amount
of innate aggressiveness and libido with which the ego is born, as well
as of the consistency of good feeding and good overall handling,
which determines the degree of frustration to which the infant is sub
jected. She pointed out both in her theoretical and case discussions
that the extent of parental gratification and good handling has a pro
found effect on the infant's development. All libidinal gratification
reduces persecutory anxiety, allowing the infant to bear frustration
more easily. Similarly, frustration exacerbates fear of the object, inten
sifying the bad-object relationship and leading to greater difficulties
in bearing negative experience. Ultimately, it is the relative strength
of the early object relationships, Klein believed, that determines the
outcome of the paranoid position.
All good feeds and good handling are initially projected onto the
good breast, just as negative experience and handling are projected
onto the bad breast. The battle for the fate of the infant's psyche is
76 Chapter 3
waged between, on the one hand, innate aggressiveness, frustrating

experience, and negative handling, all of which lead to the buildup of
the bad object, and, on the other hand, innate libido, gratifying experi
ence, and positive handling, all of which lead to the buildup of the
good object.
The Projective and Introjective Cycles
The initial projection of positive and negative experience onto the

good and bad breasts, respectively, is the first step in the projective
and introjective cycles that Klein (1952b) considered the essence of the
paranoid position. According to Klein's view, the infant reacts to all
frustrating experience by sadistically attacking the breast in fantasy,
thus endangering it. These fantasied attacks quickly become attacks
on the mother's insides as the breast is believed to have the object that
gratifies but is refusing to provide it. This early oral aggressiveness is
projected onto the breast to protect both the ego from its own destruc
tiveness and the breast from the ego's sadistic attacks. Klein believed
the fantasied attacks on the breast and the mother's insides dissipate
annihilation anxiety but produce a different threat: the infant is now
under potential attack from without, from the breast itself. The bad
object is now outside and possesses the destructiveness that once
threatened from within. Thus, the projective process transforms
annihilation anxiety into persecutory anxiety.
To reduce the danger of attack from outside, the ego introjects the
bad breast in an effort to control the danger (Klein, 1952b). Klein's
thinking is similar in this regard to Fairbairn's concept of introjection
as an effort to master the anxiety of the bad object. Introjection is the
next step in the projective-introjective cycle; like all other maneuvers
in this process, it assuages one type of anxiety but generates another.
The danger to the ego now resides within once again. It may appear
that the infant has gained nothing by shifting aggressiveness from
inside to outside and back again, but this is not the case, according to
Klein. Each step in the projective-introjective cycle does reduce anxi
ety, but its success depends on the relative balance of good- and bad-
object experience. Persecutory anxiety inevitably creates the need for
introjection, but the anxiety produced by the internalization of the bad
object is assuaged by the buildup of the internal good object from sat
isfying experiences. If there has been a strong buildup of the good
internal object, the internalized aggressiveness will not result in an
overwhelming annihilation anxiety, and persecutory anxiety will be
reduced.
Melanie Klein 77
The introjection of the bad breast is the core of the malignant, harsh
aspect of the superego (Klein, 1948a, 1958). This introjected persecu
tory anxiety is experienced by the verbal child or adult as internal ver
bal self-abuse. This view of early superego formation allowed Klein to
explain the persecutory nature of the superego in children as well as
various types of pathology characterized by psychological self-flagel
lation. She believed that all children suffer from severe superego stric
tures because the bad breast is introjected so early. Similarly, the
introjection of the good breast forms the core of the benign superego.
The balance of the good and bad introjected objects determines the
relative severity of the superego.
Klein (1948a, 1958, 1959) viewed the persistent, vicious verbal self
attacks so common in both adult and child character pathology as fix
ation in the paranoid position due to excessive introjected persecutory
objects; such self-attacks are to be distinguished from guilt over injury
to the object, which is a depressive position dynamic (to be discussed
shortly). Internalized persecutory anxiety is rooted in anxiety of attack
defended against by the introjection of the bad object. According to
Klein, if the ego is fixated at this level, the mature superego will not
develop. It should be underscored that although the introjection of
persecutory objects reduces the threat from without, it generates anxi
ety from within and thus poses a threat to the ego, the extent of which
depends on the buildup of the good internal object.
Anxiety from within can become overwhelming if the internal good
object buildup is insufficient to counteract it; now the need is not only
to expel the bad object but also to control it. The primitive ego makes a
desperate effort to control its aggressiveness by projecting the bad
object "into" the breast and identifying itself with the object; that is,
the infant attempts to control its own aggressiveness by endeavoring
to control the aggressiveness in the object. Klein (1948b) viewed this
mechanism of projective identification as the prototypical aggressive
object relationship. It should be noted that in recent years projective
identification has come to be regarded as an interpersonal process in
which the object of the projection must actually feel the affects pro
jected into him or her (Ogden, 1982). This view, discussed further
later, is an extension of Klein's concept of projective identification
as a fantasied process in which to reduce the anxiety of internal
persecution, the infant fantasies putting its self-hatred into the mother.
Projective identification, like every step of the projective-introjec
tive cycle, generates anxiety of its own. The object now becomes even
more dangerous since it contains the aggressiveness and must be con
trolled. The fate of this mechanism, too, depends on the relative
78 Chapter 3
balance of good and bad object experience. If the internal good object
has some degree of strength, projective identification may dissipate
the anxiety of attack. However, if good object experience is insufficient
and the buildup of the internal good object is weak, the infant does
not feel it can control the destructiveness of the object. To defend
against this new danger from the outside, the primitive ego uses the
only mechanism it has at its disposal: introjection of the object. This
re-introjection is now a "forceful entry" into the psyche, resulting in
the feeling of being controlled from the outside, a dangerous level of
introjection that Klein (1948b) believed was the source of paranoid
delusions of mind and body control.
According to Klein (1952b), in normal development all these steps
in the projective-introjective cycle are experienced to some degree as
the internal good object is never strong enough to eliminate com
pletely the need to reproject aggressiveness. The extent to which the
ego is forced to rely on projective identification is a major factor in the
movement of the ego toward growth or pathology. The more the bal
ance of object relations is weighted in favor of bad object experience,
the more the ego, in increasingly desperate efforts to control its own
aggressiveness, is forced to utilize the primitive mechanisms of the
paranoid position, such as projective identification and its re-introjec-
tion. Conversely, the more dominant the good object experience, the
less the ego needs to use defenses against aggressiveness and the
greater the movement toward ego integration and growth.
Although Klein (1952a) has been quite justifiably criticized for
unfounded speculation on the infant's mental processes, she felt that
the data of infant observation confirmed her views of good and bad
object experience. The terrifying screams of the small baby who has
been left were evidence for Klein that such an infant feels subjected to
attacks from a bad object. Similarly, the fact that the infant eventually
calms down when the mother returns or when it is comforted by
another meant to Klein that the infant could re-establish the good
object.
The mechanisms of good object experience are critical for ego
development. In an effort to combat the threat of the internalized bad
object, that is, to combat the dangers within caused by the introjection
of the bad breast, the good breast is also introjected. This internaliza
tion of the good object forms the core of the ego. We have already seen
that the superego is formed by the internalization of good and bad
objects. The ultimate strength of the ego and superego structures is
primarily a function of the balance of good and bad object buildup at
this early developmental phase (Klein, 1952c, 1958). Because object
Melanie Klein 79
relationships, ego, and superego structures are all intimately related,

one cannot separate object relations from mental structure. Klein
believed her views in this regard were an expansion of those of Freud,
for whom the ego was a "precipitate of abandoned object cathexes" and
for whom the superego was also an internalized object (Freud, 1923).
Splitting
The fantasied destructive attacks threaten not only the primitive,

helpless ego but also the good breast. To protect the good breast, or
gratifying object, unspoiled by aggressiveness, the infant splits the
breast into good and bad (Klein, 1937,1957). Because of the need to pro
tect the good breast from the fantasied destructive attacks, the splitting
of the object occurs simultaneously with the projection of destructive
impulses onto the breast. Splitting thus becomes a primary defense
mechanism of the paranoid position. Klein made clear that splitting is
never total. She believed that even in the earliest phase of infancy there
is some mingling of the good and bad breast in the infant's mind; how
ever, if this contact between the two objects is transitory, the good
breast remains intact. The use of splitting prevents the infantile ego
from having to experience the anxiety of injuring the good object.
The internalized good object assuages anxiety to some degree
because of the inevitable contact between the good and bad internal
ized objects, but such contact is threatening to the good object at this
phase because of its fragility. Consequently, the ego must be split into
good and bad selves. This ego splitting protects the bad self, but miti
gates the effects of the internalized good object on the internalized bad
object. Just as object splitting protects the good object, or breast, so too
does ego splitting protect the good self. Nonetheless, according to
Klein, every internalization of good object experience fosters ego inte
gration by providing a counterbalance to the internalized bad object.
The structure of the ego is a product of the internalization of good and
bad objects. As good feeds and overall good handling lead to the
strengthening of the internalized good object, the core solidity of both
the ego and superego is strengthened. To the extent that bad object
experience interferes with this development, the ego is weakened and
the persecutory internalized object influences the development of the
superego.
The result of splitting is an ego so weakened as to lack cohesive
ness. If splitting is unsuccessful in defending against bad object expe
rience because of insufficient buildup of the good object, a more
severe form of splitting will take place. To defend against persecutory
80 Chapter 3
anxiety, the good object will be exaggerated into the idealized object,
now fantasied to provide unlimited gratification (Klein, 1946, 1952d,
1957). The idealized object is identified with the ever-bountiful breast
and thereby serves as a defense against persecutory anxiety. The fan
tasied availability of this everflowing breast obliterates all frustration
and negative experience. If the idealized object disappoints in some
fashion that breaks through into awareness, the disappointment
results in the emergence of the persecutory object. Thus, Klein's the
ory makes a crucial distinction, not always made so clearly in psycho
analytic theory, between the good object and the idealized object. The
former is an inherent part of good experience; the latter is a defense
utilized only when the ego feels threatened by persecutory anxiety
and resorts to a more severe form of splitting, the cleavage between
the persecutory object and the idealized object. This degree of splitting
always reflects an excessive degree of persecutory anxiety that is not
well defended by the projective-introjective cycles alone (Klein, 1957).
The idealized object can also be introjected, and the ego will take
advantage of this opportunity to defend against dangers from within.
The idealized object is then identified with the self, and a feeling of
omnipotence results. This omnipotence diminishes annihilation anxi
ety in a fashion analogous to the way the idealized object protects
against persecutory anxiety. The fantasy of omnipotence provides the
infant with the feeling of having unlimited control over its own fate
and simultaneously obliterates awareness of all negative experience,
helplessness, and frustration. Both idealization and omnipotence,
which tend to go together, defend against the helplessness and perse
cutory anxiety of the paranoid position, resulting in the denial of all
frustration and negative experience. Thus, the ability to deny reality
becomes another primary defense in this stage of development.
The Dynamics of Envy
The helplessness and dependence of the infant in the paranoid

position makes the need for gratification aggressive. No sooner is the
infant aware of its need for the breast than it feels envious of the good
breast for having the supplies necessary for its survival (Klein, 1957).
Believing that the infant is aware very early that the source of its grati
fication is outside itself, Klein did not recognize an initial state of
merger. The very fact of the good object existing outside leads to envy
of the good breast. Further, since the bad breast is withholding, it too
is envied. To the infant, frustration always means being withheld
from, so envy is elicited in both gratification and frustration.
Melanie Klein 81
Envy goes beyond hatred to the desire to injure. The hatred of the
breast for having or withholding needed supplies produces the desire
to injure the breast. Klein believed that the desire to spoil and remove
the good object and replace it with the bad is inherent in envy. Since
envy is dangerous to the good object, it must be defended against
with the means available. The most useful initial defense against envy
is devaluation because denigration involves denial of the need for the
object. Alternatively, the object can be idealized to protect against the
recognition of the hateful desires to spoil the object that are inherent
in envy. However, since it exaggerates the good qualities of the object,
idealization can also incite envy even while attempting to defend
against it. In this case the object is devalued to defend against the ide
alization, thus providing a layered defensive structure. The child with
excessive envy may also split off the envy and become compliant,
making exaggerated efforts to please the mother.
Psychopathology
The primary defenses of the paranoid position are projection, intro

jection, projective identification, splitting, idealization, omnipotence,
and denial (Klein, 1946). All these defenses serve the purpose of
defending against aggressiveness and its associated anxieties, whether
in the original form of annihilation anxiety; its projected form, perse
cutory anxiety; or introjected persecutory anxiety. These defenses will
be utilized by all infants, but the degree to which they are employed
to protect the ego determines the propensity for the ego to become
fixated in the paranoid position.
Various types of pathology may originate from fixation in this phase
of development. If the good object buildup is virtually nonexistent, the
projective-introjective cycle ends in the desperate "forceful entry" of
the projectively identified object, which results in delusions of being
controlled (Klein, 1946). According to Klein, psychosis results from a
failure of projective identification. However, this mechanism will be
more successful if there is sufficient good object buildup to allow it to
allay anxiety. This same principle applies to projection, introjection,
denial, idealization, and omnipotence. If the internalized good object is
fragile but provides a significant portion of the ego and superego struc
tures, these defenses will become a major fixation point of the person
ality. One can identify borderline conditions, narcissistic character
disorders, and other severe character disorders as types of psy
chopathology characterized by reliance on these defenses. It should be
emphasized that, according to Klein, the entire personality need not be
82 Chapter 3
fixated at this level for paranoid position dynamics to be operative.

Neurotic patients frequently make use of paranoid position defenses
even though their personality is not organized around them.
The organization of defenses at this level implies a weak, uninte
grated ego. A stable defensive structure relying on splitting and the
projective-introjective defenses arrests the development of the ego
by preventing the integration of split ego and object states. For
Klein, the ego with adequate balance between the good and bad
objects and good and bad selves moves naturally toward the inte
gration of ego and object (the details of which will be discussed in
the context of the depressive position). However, an imbalance
arrests the ego in the early split state, preventing this movement
and inhibiting further ego growth. The result is a permanently
weak, incohesive ego that lacks the integration of internalized
objects.
Such an imbalance also arrests superego development. The inter
nalized bad objects are ruling the personality by their continual threat,
preventing the movement toward a realistic, reasonable superego
composed of the integration of good and bad objects. In lieu of an
integrated superego structure, the personality is ruled by the internal
ized persecutory object. The constant self-flagellation of the borderline
and those with other severe personality disorders is testimony to this
inhibition of superego development. The internalized good object,
which in normal development would form the core of the benign
superego, is weak and split off and has minimal influence on the inter
nalized bad object, which tyrannically rules the ego. The use of para
noid position defenses; the arrest of ego development in a weak,
incohesive state; and the failure of superego development all reflect
fixation in the paranoid position and constitute severe character
pathology.
Furthermore, according to Klein, when frustration interferes with
good object buildup, not only is the internalized good object fragile
but its instability results in greed and the desire to incorporate and
devour. When a good object is introjected, it is greedily devoured in
fantasy in an effort to protect it against potential enemies. The desire
to devour threatens the internalized good object, further contribut
ing to its fragility. The result is a cycle of greed and anxiety as the
ego desperately searches for good objects by indiscriminately identi
fying with external objects. This cycle represents Klein's understand
ing of the chameleon-like identity changes characteristic of some
types of severe pathology the most extreme form of which is the "as
if" personality.
Melanie Klein 83
As we have seen, in projective identification the external object

must be controlled in order to manage the aggressiveness projected
into it. The ego with insufficient buildup of the internal good object
feels under extreme threat and will cling to the external object in a
desperate effort to control it. If the defense works well enough to sta
bilize the ego at this level, there is no forceful reintrojection of the
external object as seen in psychosis, and the personality is character
ized by the clinging dependence so typical of borderline patients.
Thus, in Klein's view, the demands, excessive expectations, and
inability to separate, so characteristic of the severe character pathol
ogy of borderline patients, are all rooted in the desperate attempt to
control the aggressiveness projected into the other (Klein, 1946).
It should be noted that this "compulsive tie to the object" has a dif
ferent dynamic from Fairbairn's and Guntrip's understanding of
hunger for the object, as discussed in chapter 2. Klein did not see
excessive clinging to the object as evidence of a primary need unmet
in childhood, as did Fairbairn and Guntrip. It is true that she clearly
viewed early frustrations as a major source of psychopathology; how
ever, she believed that the reason the external factor is so important is
that early frustrations lead to an increase in aggressiveness, build up
of the bad object, and persecutory anxiety, thus intensifying the pro
jective and introjective cycles. Thus the "compulsive tie to the object"
is not a direct manifestation of early need but the effect of the projec
tive mechanisms initiated by the anxiety of aggressiveness, a dynamic
that sets Klein's views in direct opposition to the theories of Fairbairn
and Guntrip.
Klein was in agreement with Fairbairn and Guntrip that the projec
tive identificatory process is a schizoid mechanism and can often
result in schizoid withdrawal (Klein, 1946), but she arrived at this con
clusion in a sharply different way. When the mechanism of projective
identification dominates the personality, the external object becomes a
representation of the self. The ego is so governed by the anxiety of
controlling aggressiveness that it views the object solely as an aggres
sive threat, the aggressiveness being originally the ego's own. Since
the object is now identified with the self, such object relationships, in
Klein's view, are narcissistic. Clinging to the object potentially
intrudes upon it and blurs object boundaries, threatening such
integrity as the ego possesses. To protect the fragile ego boundaries a
schizoid withdrawal from object contact is effected; that is, in order to
ensure sufficient distance from objects, the emotional part of the per
sonality must be split off. In effect, schizoids destroy a major compo
nent of their personalities to protect a minimal sense of boundaries.
84 Chapter 3
Because this withdrawal is characteristic of fixation in the paranoid

position, Klein, acknowledging the influence of Fairbairn's views,
modified her name for this phase and called it the "paranoid-schizoid
position."
Klein acknowledged the importance of schizoid mechanisms in the
paranoid position but arrived at the concept of schizoid withdrawal
by a markedly different route from Fairbairn and Guntrip. The with
drawal from object contact, for Klein, is always rooted in aggressive
ness even though the function of the withdrawal is ultimately to
protect the fragile integrity of the ego. The pathogenic feature of frus
tration, in her view, is always the increase in aggressiveness and bad
object buildup to which it gives rise. Klein also agreed with Fairbairn
that schizoid defenses attempt to split off, even destroy, a part of the
personality, but her understanding of this effort is that only effective
withdrawal can protect the object from the patient's projected aggres
siveness. Schizoid pathology, then, in Klein's schema, is one form of
fixation at the paranoid position.
Any of the defenses against the paranoid position can become char
acter constellations if envy is so excessive that it threatens the good
object. In such cases the personality is arrested at the paranoid posi
tion, and many of the character defenses are directed against the
awareness of envy. That is to say, the dynamics of envy, according to
Klein, are a major source of severe character pathology.
The defensive layering motivated by envy of the primal object is
illustrated by Klein's (1957) discussion of a female patient who suf
fered from strong schizoid and depressive symptoms but whose
severe pathology was not recognized until she experienced some pro
fessional success. At that point in the analysis the patient had a dream
in which she was flying on a magic carpet and looking down through
a window where a cow was munching an endless strip of blanket. It
had already been established from previous dreams that the analyst
tended to be symbolized by a cow. The patient's associations indi
cated that the endless blanket represented the analyst's words, which
she now had to swallow since the patient felt she was becoming supe
rior to her. The strip of blanket reflected the worthlessness of the
analytic interpretations. The patient was shocked at this devaluing
attitude toward her analyst, whom she consciously admired.
According to Klein, this dream and other material convinced the
patient that she possessed feelings of hatred and poisonous envy
toward her analyst and wished to injure and humiliate her, forcing
upon her the awareness of a split-off, aggressive part of her personal
ity. The result of this recognition was a severe depression, as the
Melanie Klein 85
patient could not bear her destructive desires nor reconcile them with her
idealized picture of herself. "She felt bad and despicable. . . . Her guilt
and depression focused on her feeling of ingratitude towards the analyst
. . . towards whom she felt contempt and hate: ultimately on the ingrati
tude towards her mother, whom she unconsciously saw as spoilt and
damaged by her envy and destructive impulses" (Klein, 1957, p. 209).
This vignette illustrates the importance of envy in the aggressive
ness toward the primal object, as well as the layering of defenses
against it. The patient had an idealized view of herself and the analyst
in large part to defend herself against the awareness of envy. Her pro
fessional success broke through the defense by revealing a contemp
tuous attitude toward the analyst. This devaluation masked the envy
and desire to injure, feelings from which the analyst had been pro
tected. Once this protection was removed, the patient fell into a severe
depression, convinced she had injured the analyst whom she relied
upon—and, ultimately, her mother, the primal object whom she felt
she had irreparably injured with her envy.
The foregoing describes Klein's view of character pathology orga
nized around defenses against envy. If the envious threat to the good
object is extreme, however, envy interferes with its introjection, and
the ego, now lacking the internalized good object, is further weak
ened. A weak ego tends to envy any potentially good object all the
more. A downward cycle takes place in which envy blocks introjection
and ego growth and the arrested ego becomes increasingly envious.
Further, the intent to destroy the good object leads to premature guilt,
which the ego is too undeveloped to manage and which must be pro
jected. This early form of guilt is a sadistic attack on the ego. Its projec
tion shifts the source of attack to the outside, leading to persecutory
anxiety. In this way, the object becomes a persecutor. Thus, according
to Klein's formulation, envy is a primary source of paranoia; this fact
explains why guilt is so often confused with persecutory anxiety
in the paranoid patient. When guilt is prematurely generated in the
paranoid position, it is so burdensome to the ego that it can only be
projected, resulting in persecutory anxiety.
These dynamics of the paranoid position are illustrated in Klein's
discussion of an obsessional neurosis in a six-year-old girl, Erna
(Klein, 1946). The little girl suffered from sleeplessness precipitated
by her fear of robbers; obsessional activities, including head banging,
rocking, thumb sucking, and masturbating; depression; and a severe
learning inhibition. She also dominated her mother, whom she
watched over compulsively, and she felt responsible for all her
mother's illnesses and expected punishment for them. Erna had
86 Chapter 3
shared her parents' bedroom and witnessed the primal scene. In her
therapeutic play Erna provided abundant evidence of her hatred and
envy of both parents. For example, she played the mother and had
the analyst suck on two "red, burning" lamps, which she then put in
her own mouth. (Klein interpreted the lamps as mother's breast and
father's penis). The play was always followed by "attacks of rage,
envy, and aggression against her mother, to be succeeded by remorse
and by attempts to make amends and placate her" (Klein, 1946, p. 68).
Erna also played intensely rivalrous games with the analyst in which
she came out ahead. She cut paper and said that "blood was coming
out." Erna played at being a washerwoman, a role in which she pun
ished and humiliated a child. Once she changed from a washerwoman
into a fishwife; she turned on the water tap, wrapped paper around it,
drank greedily from it, and chewed an imaginary fish. Klein (1946)
believed this play material showed "the oral envy which she had felt
during the primal scene and in her primal phantasies" (p. 70).
Erna also played at cheating the analyst, but since she had a police
man on her side, the analyst was helpless against her. She also showed
her hostility to and envy of her mother by pretending to be a queen or
a performer admired by spectators while the analyst, in the role of a
child, was mistreated and tormented. While the patient and her hus
band ate delicious food, the child had gruel and was made sick. The
child was made to witness sexual intercourse between the parents and
was beaten if she interrupted them. In one game a priest gave a per
formance and turned on the water tap; his partner, a woman dancer,
drank from it while the child, named Cinderella, could only watch
and remain motionless. At this point Erna let loose a rage attack that
"showed with what feelings of hatred her phantasies were accompa
nied and how badly she had succeeded in dealing with those feelings"
(Klein, 1946, p. 72). Every educational measure, every limit set upon
her, was interpreted by Erna as her mother's sadism. The little girl,
feeling persecuted and spied upon by her mother, was terrified of her.
Klein interpreted the child's anxiety, sleeplessness, and fears of rob
bers as rooted in her hostility and envy toward the mother and her
desire to damage her insides. Erna's destructive impulses were pro
jected onto the mother, resulting in persecutory anxiety that was
displaced onto fears of robbers, thus leading to sleeplessness. Her
sadism and envy were also projected into the mother via projective
identification, leading to the need to maintain constant vigilance over
her. Thus, the control the little girl attempted to exert over her mother
was, in Klein's view, a product of the projective identification of her
aggressiveness and of her sadism and envy.
Melanie Klein 87
Klein viewed Erna's pathology as so rooted in overwhelming

aggressiveness that she did not consider the child's witnessing of the
primal scene as pathogenic in itself. She believed that the child wished
to do what her mother did with her father and that because she could
not, the observation of the primal scene led to an increase of aggres
siveness, envy, and sadism. The primal scene became pathogenic
because the child's ego could not cope with her destructive impulses
toward her mother which led to the projection of her aggressiveness
and sadism both onto and into her. Further, Klein interpreted the
obsessive thumb sucking as rooted in fantasies of biting and de
vouring the mother's breast. The learning inhibition, in Klein's view,
resulted from the association of writing and arithmetic with sadistic
attacks on the mother's body. In brief, Klein believed that the child's
overw helm ing envy and aggressiveness toward her m other
constituted the crucial dynamic underlying all her symptoms.
The case of Erna illustrates Klein's understanding of pathology
originating in the paranoid position as well as her view of neurotic
cases: she believed neurosis is rooted in the psychotic anxieties of the
paranoid and depressive positions. Erna's primary defenses were pro
jection and projective identification, and her anxiety was clearly perse
cutory in nature. Klein believed this case illustrated the pathogenesis
of excessive aggressiveness and envy, which can dominate the ego
and lead to its reliance on the paranoid defenses of projection and
projective identification.
The good object acts as a counterbalance to envy and hate, just as it
counterbalances the internalized bad object. Good experience pro
duces good internalized object relationships and enjoyment; a natural
product of the enjoyment is gratitude. The degree of enjoyment expe
rienced in the paranoid position expands the capacity for gratitude
later on, whereas, excessive aggressiveness and arrest in the paranoid
position inhibits the development of the capacity for gratitude. The
interference of intense envy and destructive impulses with enjoyment
and gratitude explains the lack of a capacity for gratitude in severe
character pathology.
A variety of severe character pathological constellations may origi
nate in the paranoid position. While a complete deterioration of the
defensive system results in delusional paranoia, more frequently,
unresolved anxieties from the paranoid position are arrested at a more
advanced defensive level. If the projective-introjective cycles allay
anxiety at the level before the deterioration to delusion, the personal
ity is fixated at the level of projective identification, and the result is
severe character pathology. Fixation at the nonpsychotic paranoid
88 Chapter 3
position involves the following characteristic defenses: idealization,

omnipotence, denial, devaluation of the object, and splitting. An
ego structure reliant on these defense constitutes severe character
pathology tantamount to what is now commonly referred to as the
borderline personality. Conditions now typically labeled narcissistic
personality disorders are conceptualized in a similar fashion, with
emphasis on the idealization and omnipotent defenses and the use of
schizoid mechanisms to regulate ambivalence in object relationships.
As we shall soon see, some of Klein's followers have formalized these
views into more precise conceptualizations of narcissistic and border
line pathology (Rosenfeld, 1971, 1978; Segal, 1983). As is discussed in
chapter 5, Kernberg's (1975) formulation of and treatment approach to
borderline and narcissistic personality organizations are based largely
on Klein's concepts. If the schizoid mechanisms become the dominant
force in the personality, the emotional component of the personality is
buried and a schizoid personality is the result.
The commonality of all these forms of psychopathology is that the
ego feels endangered by the projection of its own aggressiveness. The
ego in its primitive, incohesive, and fragile form is always a victim,
experiencing itself as ill equipped to mount a defense against the pow
erful, hostile forces it feels subjected to. When the ego begins to move
toward the experience of itself as an agent, that is, to develop a sense
of responsibility, a new developmental phase is initiated.
The Depressive Position
If early positive experiences and innate libido have been strong

enough to result in a solid internalized good object, the growing ego
will be able to begin to recognize that the good and bad objects are the
same. Although some contact between good and bad objects occurs in
the paranoid position, Klein (1937) believed that a new developmen
tal phase is initiated when the infant ceases to use splitting as a major
mechanism for the organization of its object relationships and begins
object integration. Klein believed that this shift begins at about three
to four months and is completed at about six months, when, she
believed, the oedipal phase is initiated. The integration process
appears to be conceptualized primarily as a developmental unfolding
given satisfactory early experiences and innate libido. In Klein's view,
however, this movement is also motivated by the anxieties of the
paranoid position. The most positive response to persecutory anxiety
is to begin object integration.
Melanie Klein 89
When the infant becomes aware of the fact that the object it hates
and desires to injure and destroy is the same object it loves and
depends on for gratification, it has begun to move from the "part
object" level of object relations to the experience of whole objects
(Klein, 1935, 1940). Simultaneously, the infant begins to remove itself
from the painful role of victim, a role characteristic of the paranoid
position; it now feels its desire to injure the object of its love. This
recognition enables the growing ego to experience a feeling of agency,
the power to injure. Rather than experiencing itself as the passive vic
tim of persecution, the infant believes itself to be the agent of injury;
the persecution is now of the object rather than the ego, and the source
of danger is within rather than without. Klein called the realization
that one can injure the object of its love the"depressive position" and
the anxiety that this feeling engenders "depressive anxiety."
According to Klein (1948a), guilt originates from the anxiety of
injuring the loved object. Because integrated object perception stimu
lates the recognition of the desire to injure, whole object perception is
inevitably accompanied by guilt and anxiety. This view represents a
major disagreement between Klein and followers of classical psycho
analytic theory. Klein believed that guilt stems from destructive wishes
toward the loved object, rather than from sexual longing, and that it
appears long before the consolidation of the oedipal phase at about
three years of age. Moreover, Klein believed that because guilt arises
in this way it is closely linked to the anxiety of loss, as the ego fears for
the loved object, and inevitably results in the reparative desire. Guilt,
then, is the bridge between the destructive desire and reparation.
The movement toward object integration is halting and oscillatory,
since whole object perception inevitably gives rise to depressive anxi
ety, resulting in the urge to regress to the paranoid position. Some
movement backward is an inevitable component of the lengthy
process of object integration, but if good object internalization is well
established, the infant will be able to sustain whole object integration.
Despite the chronological priority of the paranoid position, Klein
(1937) often seemed to regard the depressive position as the funda
mentally more important developmental phase. She tended to empha
size it in her discussions of psychopathology and, in fact, formulated
it before she conceptualized the paranoid position. Indeed, the distinc
tion between these two critical developmental phases is not always
clear: envy presumably originates in the paranoid position, and yet it
involves the desire to injure the gratifying object. Klein did not seem
to regard envy in the paranoid position as dependent on the object
integration of the depressive position.
90 Chapter 3
If the infant is able to feel in its fantasies that it has repaired and
restored the injured object, guilt will not become crippling. Reparative
experiences are crucial to overcoming the burdensome guilt anxiety of
this phase. The capacity for repair is dependent both on the introjec
tion of the good object in the paranoid position and on continued
good handling, which allows for the perception of good external
objects. If these conditions occur, the positive experiences provide the
child with the opportunity to feel that its aggression has not irrepara
bly damaged its loved object, a feeling that mitigates its guilt.
However, if the introjected good object is weak and unstable or the
experiences in this phase are predominantly negative, the infant's per
sonality will tend to become dominated by excessive unresolved guilt
and feelings of unworthiness. (As will be seen in chapter 4, the con
cepts of ambivalence toward whole objects and repair to mitigate the
ensuing guilt greatly influenced Winnicott's developmental theory
and treatment approach to character pathology.)
Klein believed that the ability to love and sustain love relationships
in later life is dependent on the experience of reparation in the depres
sive position. Without this experience the personality remains chroni
cally fearful that its aggressiveness will injure or destroy the loved
object. The result is constant uncertainty regarding love relationships,
especially when aggressive feelings appear. For this reason, Klein
believed that good object relationships in adult life depend to a large
extent on the outcome of the depressive position. She also believed,
however, that successful love relationships in later life can help to
complete unfinished reparation from the depressive position (Klein,
1937). The problem with this later life resolution is that if reparation is
blocked completely, successful love relationships are not possible. The
reparative experience in the infantile depressive position must be suf
ficiently complete in order for later love relationships to develop to
the point where they can resolve arrested infantile efforts at repair.
Parental love and the provision of good care eases the child's anxi
ety of its destructive fantasies and aggressiveness by proving that they
have not destroyed the good mother (Klein, 1937). All good experi
ence in this phase, according to Klein, proves to the child that its inter
nal fantasies of destruction have not been carried out in reality,
rendering the child's feelings of danger to the object less severe. All
negative experience increases the child's aggressiveness and belief in
its power to injure. If experience in this phase is excessively frustrat
ing, the child will feel that its aggressiveness has in fact destroyed the
loved object, resulting in intolerable guilt. However, parental care is
not the only variable. If the child possesses extreme innate aggressive
Melanie Klein 91
ness and minimal ability to tolerate frustration, its perception of the

parents will be dominated by aggressiveness even if they love the
child, and the child will feel that much more burdened by excessive
guilt. Similarly, some children receive bad treatment and continue to
perceive the parents positively because they do not have a great deal
of innate aggressiveness to project onto parental figures. In this case,
guilt is minimal despite mistreatment.
Awareness of the fantasied ability to injure the loved object
increases envy and greed (Klein, 1957, 1958). As the child begins to
feel the anxiety of losing the loved object, it feels less certain of it and
feels more greedy for and envious of the good object it cannot secure.
As envy and greed increase, fear of destroying the good object is exac
erbated, leading to more greed and envy, and a pathological cycle
ensues. The counterbalance to this negative cycle is the internalization
of good objects and reparation, both of which allay the anxiety of los
ing the good object. If positive experiences are not sufficient to coun
teract the pathological cycle of envy and depressive anxiety,
emotional organization becomes fixated at the depressive position and
a pathological result ensues.
Defenses in the Depressive Position
Because the anxiety of losing the loved object is so painful and the
consequences so potentially disastrous, defenses are erected against
depressive anxiety. The infantile ego is so weak that to manage the
overwhelming anxiety of intending to injure the loved object, it resorts
to the omnipotent defense of fantasizing restoration of the parental
figure. Because the reparative experience includes the omnipotent
belief in the magical ability to control bad objects and restore good
objects, mania is a normal accompaniment of the depressive position
(Klein, 1940). In Klein's view, omnipotence is the fantastic belief in
absolute control and therefore involves denial of psychic reality.
Aggressiveness, bad objects, and dependence on real objects are
denied, and good objects, believed to be under omnipotent control,
are idealized. The denial of psychic reality is therefore an inevitable
component of the infant's need to master the depressive position.
Excessive depressive anxiety is, therefore, defended against by
massive denial of the dangers to the good object (Klein, 1935). In the
manic state all dangers to the good object disappear and the object is
magically restored. Thus, the importance of good objects in reality is
denied, as is all object danger. In this way, all guilt and dread disap
pear. Thus, the manic defense, for Klein, is the ego's denial of both
92 Chapter 3
external and psychic reality and the flight to the exaggerated internal
ized good object. The manic state so massively defends against the
anxiety of losing the loved object that all dependence on objects is
denied and the only object relation is to the internalized, idealized
object with which the ego now identifies itself.
In the normal developmental process frustration and negative
experience inevitably occur, forcing reality upon the child to some
extent. Consequently, fluctuations between the manic and depressive
positions are continuous throughout this period as the infant alter
nately denies and perceives psychic reality (Klein, 1935, 1940). When
the manic defense fails, the child is forced into reliance on obsessive
mechanisms in a desperate effort to repair the object over and over
again to prevent psychic disintegration. In Klein's (1935) view, this
mechanism constitutes the origin of childhood obsessional symptoms.
Because the early aggressive intent was not successfully repaired, the
child is attempting to control magically the object it fears it has
destroyed. Since the manic defense never works perfectly, some
degree of obsessional behavior is an inevitable part of childhood. The
adult obsessive is fixated in this endless effort to repair magically the
fantasied injury to the loved object.
Psychopathology
Unresolved depressive anxiety leads to a defensive arrest that may

assume a variety of pathological forms. If reparative efforts are not
felt to be successful, the continued need for reparation may appear in
a desire for perfection. Work inhibitions, in this view, are rooted in the
fear of imperfection that results from the need to reassure oneself that
one has not irreparably damaged the loved object. The obsessive
worker who cannot commit to a project for fear of committing an error
is, in Klein's view, experiencing the work project as an opportunity to
repair the damage to the loved object, but an opportunity doomed to
failure since only a perfect product can relieve the guilt. If guilt is so
overwhelming that no reparation is considered possible, it will be
massively repressed. In this case, the superego is crushed and the
source of damage is externalized, resulting in sociopathy (Klein, 1933).
Criminality, in Klein's view, is a product not of a gap in the superego
but of an excessively burdensome superego that is repressed.
If the child does not feel it can repair the damage to the good object,
the ego may become fixated in the manic defense. The mania of the
adult manic-depressive is, according to Klein (1935) the equivalent of
the normal child's temporary defense against depressive anxiety. If
Melanie Klein 93
the child feels its aggressiveness toward the good object is so danger
ous that its very existence is threatened, the ego will continue to deny
all psychic reality. The result is an ego reliant on the manic defense,
and when depressive anxieties are aroused in adulthood, the result
will be manic-depressive illness.
These views have clear implications for the mourning process.
Klein (1940) agreed with Freud's (1917) view that the task of mourn
ing is to introject the lost object. However, Klein departs from Freud in
her view that the loss evokes depressive anxiety. According to Klein,
the infantile fantasy of having injured or destroyed the parents is
inevitably activated by adult loss. Thus, the outcome of mourning will
depend on the degree of successful resolution of the infantile depres
sive position. If the infantile object was felt to be repaired, when the
infantile depressive position is activated by adult loss, the mourner
will be able to repair and restore the newly lost object internally by
restoring the injured parents from infantile fantasy life. The reparation
of the early objects and their reinstatement allows for overcoming
grief and making peace within. If, on the other hand, there is unre
solved guilt for having damaged or destroyed the parents in fantasy
and the early objects are not repaired, the mourner feels that he or she
has once again destroyed the loved object. Just as he or she did in the
infantile depressive position, the mourner fears retaliation for the
injury, and fears of punishment and persecution become part of the
grief reaction. In this situation, the mourner becomes paralyzed by the
guilt of destroying the lost object in fantasy and by the ensuing anxi
ety of retaliation. The guilt is too much to bear and the outcome is an
inability to overcome the grief of mourning. Since the reinstatement of
good objects does not occur, when adult loss is responded to in this
way, mourning becomes melancholia and results in the same efforts
to escape depressive anxiety as are found in melancholia: mania,
obsessional defenses, paranoia, and "flights to the object."
The most typical pathological outcome of unresolved depressive
anxiety is a chronic fear of injuring the loved object that results in clin
ical depression (Klein, 1935,1937,1940). The melancholic, according to
Klein's model, has been unable to overcome the anxiety of damaging
its loved objects and, consequently, has repressed all aggressiveness.
The relentless internal self-persecution of the depressive, according to
Klein, is a product of its self-hatred for having injured the loved
object. Not having the opportunity for reparation, the melancholic
attempts to save the object the only way it can; namely, by turning its
aggressiveness toward itself. The self-abuse of the depressive is not
only a product of guilt but also a desperate effort to protect the good
94
object. All enduring relationships are subject to rupture and even ruin
because the individual feels, however unconsciously, that he or she is
dangerous to the early object of its love. As soon as gratification from
an object becomes possible, the depressive unconsciously fears injur
ing or destroying the object of its love. If such gratification does occur,
the individual feels guilty, however unconsciously, and a depressive
episode will ensue.
According to Klein (1937), unresolved guilt from the depressive
position also explains why some depressed patients become so hope
lessly attached to their objects that they have great difficulty separat
ing from them. Unconscious anxiety from potential destructiveness to
the loved object requires constant contact as reassurance that the
destruction has not taken place. By attaching tenaciously to the object,
such patients have found a way to maintain object contact despite
unresolved intent to destroy. The price of this quasi resolution is the
endless reassurance needed to maintain the certainty of the object's
preservation and the continuation of the relationship. This type of
excessive object attachment is to be distinguished in Klein's work
from the "compulsive tie to the object" based on persecutory anxiety
as illustrated in the treatment of Erna. According to Klein, all difficul
ties in separation from an external object originate in aggression
toward the object and are therefore motivated by either persecutory or
depressive anxiety. This view is in clear disagreement with the
Fairbairn-Guntrip position (discussed in chapter 2) that excessive
object attachment is a deprivation-induced arrest of the normal devel
opmental needs to depend on a reliable object. On the other hand,
Klein's view that extreme dependence is a defense against hostility to
the object at either the depressive or paranoid levels had a profound
influence on Kernberg, who (as we shall see in chapter 5) based his
conceptualization of severe character pathology on the pathogenicity
of excessive aggressiveness.
How far Klein (1960) extended this concept of separation anxiety
can be seen in her treatment of Richard. Because she treated this ten-
year-old boy during a planned four-month-stay in a small Scottish
town during the war, the analysis had a prearranged ending. As the
time for Klein's return to London and termination of the analysis
drew near, Richard began to react to the impending separation.
Richard was staying with a man, Mr. Wilson, whom he felt was strict.
In the 76th session, Klein had some oranges in a parcel, and Richard
"looked white with anger and envy, but said he did not like oranges"
(p. 388). When later in the session Richard begged Klein to take him
with her and let him sleep in her bed, she interpreted "Richard felt
Melanie Klein 95
that because Mrs. K. did not let him stay with her, give him the
oranges, and love him, she had become the ally of Mr. Wilson, who
was now felt to be the bad Daddy. This made his need for reassurance
and love from her all the greater" (p. 389). Richard was restless and
depressed throughout the session and at one point talked of a tornado
razing two houses to the ground. Later in the session Klein inter
preted that his depression was due to her going away and to his jeal
ousy of her grandchildren and the other patients she would see in
London. She told him he wanted to raze her house because he was
angry with her for leaving and "therefore his fear of losing her forever
was very great" (p. 390). This vignette demonstrates the emphasis
Klein placed on aggressiveness and the fear of destroying the loved
object in the dynamics of separation anxiety, an emphasis she main
tained even when the loss was real and imposed from the outside. In
this case, as opposed to that of Erna, the interpretation was focused on
the anxiety of loss due to aggression toward the loved object.
The easiest path of escape from the depressive position is regres
sion to the splitting of the paranoid position, (Klein, 1937). When the
good object is not established well enough within, depressive anxiety
is excessive and the defense of least resistance is to resplit the object to
immediately allay the anxiety of loss. This regression evokes the per
secutory anxiety of the paranoid position, rendering the infant help
less once again; its need to gain control of its persecutors will now
motivate a renewed advance to the depressive position. This oscilla
tion is Klein's explanation for why so many patients cycle between
paranoia and depression. The target of their aggressiveness continu
ally shifts between the object and the self as they oscillate between
forward movement to the depressive position and regression to the
paranoid position.
Escape from depressive anxiety may also be effected by denial of
dependence on the object. At the neurotic level the individual turns
away from all loved objects, fearing dependence on them (Klein,
1937). Consequently, the neurotic pattern of inability to commit to a
partner and remain in a relationship is always at root the unconscious
fear of injuring the loved object. This theoretical outlook provides a
slant on this neurotic syndrome, seen so much in today's clinical prac
tice, that is different from the common conception of narcissistic vul
nerability (discussed in the context of Kohut's theories in chapter 6).
Indeed, Klein (1937) viewed the fundamental dynamic of the Don
Juan syndrome, or any type of chronic infidelity or promiscuity, as the
need to reassure oneself that one is not dependent on one object and
therefore not in danger of doing it harm. Klein believed the source of
96 Chapter 3
dependence anxiety lay in the identification of the current loved object

with the original loved object, which in fantasy has been irreparably
injured by the aggressiveness directed toward it. In some cases the
outcome is an avoidance of affective life as a whole; the individual
with such a constricted personality has renounced all affective bonds
in order to avoid reexperiencing the original injury to the object.
The "flight to the external object" as a means of escaping the
depressive position can also occur in milder forms. The guilt of the
depressive condition, according to Klein (1937), is the primary reason
for inferiority feelings. Unconscious hatred of the loved object in the
infantile position makes the adult feel unworthy of being loved. This
neurotic level of feelings of unworthiness and low self-esteem is to be
distinguished from the internal persecution of the paranoid position
and is characterized by a continual need for praise and admiration
from external objects to assuage the guilt and low self-esteem that
result from the unconscious intent to injure the original loved object.
The need for esteem from the environment may assume the form of
excessive concern over competence, beauty, work, or many other spe
cific areas of life. For Klein, the root of all such self-esteem anxiety is
unconscious guilt, and the effort to assuage this guilt constitutes the
neurotic outcome of unrepaired desires to injure in the depressive
position. This excessive dependence on the object's opinion is neu
rotic, rather than more severely pathological, because the need for
admiration tends to be specific, rather than all-consuming and the
dependence is on the other's admiration, rather than a clinging object
tie. According to Klein (1937), the determining factor of whether
dependence on others is neurotic or more severely pathological is the
timing of the trauma and guilt fixation. If the fantasy of damage to the
loved object occurs in the early part of the depressive position, when
only a minimal degree of integration has occurred, anxiety of injuring
the loved object will be pervasive and dependence will tend to be
extreme. However, if the ego has achieved substantial whole object
and ego integration, the fear of damage will be less extreme and the
need for reassurance from the object will apply only to a specific area
of functioning. In either case, the dependence on the object rooted in
the depressive position is to be contrasted with overdependence origi
nating in the paranoid position, which, as discussed earlier, is moti
vated by the infant's fear of the bad object and need to cling to the
good object for protection.
Other forms of psychopathology can be rooted in either depressive
or paranoid anxieties (Klein, 1935). For example, eating disorders can
be due to the identification of food with persecuting objects, in which
Melanie Klein 97
case food is experienced as a poisonous attack on the body, or with

the good object, which is endangered by being taken in. The fear of
destroying or injuring the good object by biting or chewing is depres
sive anxiety displaced onto food whereas the anxiety of being dam
aged by the ingestion of external substances is the paranoid type.
Similarly, hypochondriasis can be symptomatic of either type of anxi
ety. If the body is felt to be under attack, bad objects taken from the
outside are felt to be persecuting the body. If, however the hypochon
driacal symptoms result from internal warfare in which good objects
are under attack from internalized bad objects, the anxiety is depres
sive. In both cases the result is preoccupation with danger to the body,
but in the first case the danger is felt to come from outside persecutors
and in the second type the danger is to the object experienced as
within. The comparison between the roots of these two clinical syn
dromes, eating disorders and hypochondria, illustrates the essential
distinction between persecutory anxiety and depressive anxiety: the
former is danger to the ego and the latter is danger to the object. In
almost all of Klein's clinical cases there is some degree of mixture of
the two, (although one type of anxiety will tend to predominate in a
given case), and within a particular session or theme either depressive
or persecutory anxiety will come to the fore.
Klein's (1935) views on fixation in the depressive position and its
relationship to hypochondria are illustrated in her discussion of patient
X. The analysis of this case is discussed more fully later; the part of
the analysis relevant to this discussion occurred in the patient's move
ment from paranoid to depressive anxiety. When this change took
place, X became deeply depressed and his hypochondriacal pains
shifted: in the first phase of treatment the analyst had been identified
with a fantasied tapeworm and other substances attacking his insides;
after the shift to the depressive position X had a fear that cancer would
eat away through his stomach. But X wanted to protect the analyst—
identified with the internalized mother—from attack by his own
sadism and greed, which he equated with the cancer and in conse
quence of which he felt despairing, unworthy, and deeply depressed.
When paranoid anxiety was predominant, X was hypochondriacal, but
he felt attacked rather than concerned for the object. According to
Klein, when X shifted to the depressive position, he no longer felt
attacked but now feared that his illness, a cancer, would injure the ana
lyst; in consequence of this fear, the patient became depressed.
The dynamics of fixation in the depressive position are also well
illustrated in Klein's (1935) discussion of two dreams of patient C, who
also suffered from severe depression, paranoia, and hypochondria. In
98 Chapter 3
the first dream C's parents were elderly, and he was "managing"
(taking care of) them on a trip in open air. The parents were lying in
bed, their beds end to end, and the patient found it difficult to keep
them warm. With the parents watching, the patient urinated. He
noticed that his penis was very large; he felt uncomfortable because
he believed that he would be humiliated if his father saw the size of
his penis. However, he also felt he was urinating to save his father
the trouble, and he commented that he felt as if his parents were a
part of himself. The next night C dreamed that he heard a frying
sound. He felt that a live creature was being fried. He tried to ex
plain to his mother that to fry something alive was the worst thing to
do, but she did not seem to understand. He associated to beheading
and acknowledged that he used to think about torture.
According to Klein's (1935) interpretation, the urinating repre
sented the early aggressive fantasies toward the parents, especially
toward their sexual relationship: "He had fantasied biting them and
eating them up, and among other attacks, urinating on and into his
father's penis, in order to skin and burn it and to make his father set
his mother's inside on fire in their intercourse. . . . Castration of the
father was expressed by the associations about beheading" (p. 281).
Klein added that the patient's wish to humiliate the father was shown
by his feeling that he ought not to do so. His sadistic fantasies were
represented by his mother's inability to understand that she was in
danger from the biting penis inside her. Klein interpreted the position
of the beds to indicate both C's aggressive and jealous desire to sepa
rate his parents during intercourse and anxiety that sexual contact
would injure or kill them, as was his wish. C now felt overwhelming
anxiety that his parents would die.
There is much more material in C's dreams, but this brief account
shows that his dominant anxiety was distress and concern for the
loved object, the danger C imagined them to be in coming from his
own aggressive desires toward them. Klein (1935) comments:
The patient deals with the depressive position in different ways. He uses
the sadistic manic control over his parents by keeping them separated
from each other and thus stopping them in pleasurable as well as in dan
gerous intercourse. At the same time, the way he takes care of them is
indicative of obsessional mechanisms. But his main way of overcoming
the depressive position is reparation. In the dream he devotes himself
entirely to his parents in order to keep them alive and comfortable [p. 283].
These dreams are illustrative of the way Klein believed the mecha
nisms of the depressive position work. In order to love his parents, C
Melanie Klein 99
had to repress his sadistic desires toward them and repair the loved
objects in fantasy, as represented in the first dream by his over-solici-
tousness and caretaking of elderly parents. Because the patient felt
overwhelming guilt for these desires, he had a need to repair and con
trol his parents, which resulted in the overattached "management" of
them in reality.
In summary, Klein was able to explain a great deal of psycho
pathology as fixation at issues embedded in the depressive position.
Not only did she view depression itself as originating in this develop
mental phase, but she conceptualized many other manifestations of
psychopathology as attempts to escape from depressive anxiety. In
some cases the link with depressive position dynamics is evident, as
in mania and unresolved mourning, but the depressive position
dynamics of other types of pathology are not so obvious. Klein also
conceptualized obsessive neurosis, some forms of paranoia, neurotic
character pathology, and many of the symptoms in cases that today
would be labeled borderline or narcissistic as rooted primarily in the
drive to escape the anxiety and guilt of the depressive position.
Because depressive position dynamics imbued so many types and lev
els of pathology, Klein gave the depressive position the central role in
her theory of pathogenesis and seemed frequently to speak of it as the
fundamental conflict of childhood even to the point of equating it
with the infantile neurosis. According to Klein, the extent of success in
overcoming the anxieties of this fundamental conflict is the most criti
cal feature in emotional development and in preparing the child for
oedipal conflict.
The Oedipus Complex
As we have seen, Klein believed that the superego originates in the

early introjective process and that guilt originates in the depressive
position. This distinction would seem to set her views in opposition to
Freud's belief that the Oedipus complex is the source of guilt and the
superego. However, instead of opposing Freud in this way, Klein
adopted the position that the Oedipus complex begins long before the
origin ascribed to it by traditional psychoanalytic theory (Klein, 1928,
1933,1945). Klein (1926) believed that observations of children and the
findings of child psychoanalysis, of which she was a pioneer provided
"evidence" for the existence of guilt, the superego, and the Oedipus
complex in children much younger than three years of age. Klein
pointed to the fact that children show preference for the parent of the
opposite sex as early as the beginning of the second year as evidence
100 Chapter 3
that the Oedipus complex originates at least that early. In children as

young as two-and three-quarters to four years old, Klein (1933) saw
evidence of early, harsh superego strictures in their fantasies of crea
tures that bite, devour, and attack. It was Klein's contention that only
the existence of oedipal ambivalence and rivalry could account for the
severity of the superego and the accompanying fear of parental retri
bution in children this young and that the "full-blown" Oedipus com
plex in the fourth or fifth year is therefore a result of a process begun
in infancy.
By placing the origins of the Oedipus complex in the period of the
projective-introjective cycles, she was able to equate early persecutory
anxiety with the maternal retribution of the oedipal phase. In Klein's
view, by the beginning of the second year sadistic impulses and perse
cutory anxiety are related to oedipal rivalries. This contention links
persecutory anxiety, the early sadistic superego, and the origins of the
Oedipus complex.
According to Klein, the oedipal phase begins with the height of the
depressive position in weaning when the infant searches for a new
object; the second object, in Klein's (1945) view, is the father's penis. If
the breast was gratifying, the good object is sufficiently internalized
that new objects will be sought at this point. If the breast was unduly
frustrating, the infant will turn to new objects to escape the frustra
tion. According to Klein, the breast and penis in this phase are both
oral objects and both are split into good and bad. To the degree that
the good breast has been solidly introjected, the penis will be viewed
as good; likewise the strength of the bad breast will determine the
influence of the bad penis. One can see that Klein did not view libidi-
nal positions as discrete, sequential phases of development. She
believed that the introjection of objects begins to become genital as
early as six months of age, although the organs are oral. Similarly, she
believed that the aggressiveness of this phase quickly becomes anal
sadism as the infant desires to expel introjected objects. Oral, anal, and
genital aggressiveness intermingle as the oedipal phase dawns. Envy
plays an important role in the beginning of the oedipal phase as both
good breast and bad breast envy motivate the infant to turn away
from the breast to search for the penis.
This view raises the question of how an infant could have knowl
edge of the sexual organs and attribute meaning to them. Klein treated
many children who had witnessed the primal scene, and she attrib
uted their knowledge of sexuality to this observation. Although she
realized that not all disturbed children have this experience, she felt
that the content of their play indicated unconscious awareness of sex
Melanie Klein 101
ual intercourse. Klein (1946) was forced to adopt the hypothesis that
children are born with innate knowledge of sexual intercourse.
All that has been said thus far regarding the Oedipus complex is
true for both sexes. However, the boy and girl experience the shift
from breast to penis differently (Klein, 1946c). The little girl wants her
father's penis but fantasizes that her mother has it, a fantasy that initi
ates rivalry with her mother and fantasies of attacking her mother's
body and stealing its contents. In fear of retaliation, the little girl turns
to the father. In Klein's view, the little girl wishes to have a penis to
escape maternal retaliation; thus penis envy is a product of oedipal
conflicts, not a cause. If the aggressive feelings toward the mother are
not excessive and the good object has been sufficiently internalized,
the little girl will eventually identify with her mother as she realizes
that her wish for her father's penis is futile. However, if she has exces
sive oral sadistic wishes toward her mother from earlier phases, she
may not be able to overcome her fear of maternal retribution and will
not be able to enter the oedipal phase or, at best, will be unable to
resolve her oedipal rivalry. Her wish for the father's penis may
endure, or if given up, it will result in regression, rather than identifi
cation with the mother. Thus Klein's contribution to oedipal theory is
that, unless the paranoid and depressive position resolution has mod
ulated the little girl's aggressive wishes toward her mother, she will
have great difficulty even entering the oedipal phase, much less mas
tering it. For example, if the girl feels she has injured the mother, she
may have to disavow her aggressiveness toward her mother and
regress to splitting, a sequence that leaves her unable to enter the
oedipal rivalry and arrests the personality in the preoedipal phase.
Klein's (1945) patient Rita is a prime example of a little girl so trau
matized by her unresolved aggressiveness toward her mother that she
was unable to master the oedipal phase. Rita was two-and-three-quar-
ters years old at the onset of her analysis and already suffered from a
variety of symptoms—anxiety, appetite disorder, depression, obses
sional symptoms, and inability to tolerate frustration. She had great
initial difficulty accepting the bottle and later resisted yielding the bot
tle for food. She had witnessed the primal scene and felt that her
father was sadistically damaging her mother. At the beginning of her
second year, she switched her preference from mother to father and
tried to exclude mother. At 18 months, her mother became her
favorite, and at that time Rita developed phobias and nighttime ter
rors and began clinging to her mother so intensely that she could not
separate from her. Klein pointed out that Rita's relationship with her
mother was dominated by persecutory fear and depressive anxiety.
102
She feared her mother, who was her loved and needed object but
whom she "endangered" with her aggressive wishes. Rita was cling
ing to her mother to protect the needed object from her own aggres
sive wishes. Consequently, when Rita began the oedipal phase at the
beginning of her second year, she could not tolerate the anxiety of the
rivalry and regressed to the helpless, clinging dependency of earlier
infancy. Nor could she really desire her father, for she felt he had
sadistically damaged her mother and was therefore a threat to her, a
fear that also fostered her regression.
This case could be discussed from many perspectives, the point in
the current context is that Rita's unresolved destructive wishes toward
both parents, but especially her mother, rendered the Oedipus com
plex insurmountable. Klein believed that the clear shift in Rita's pref
erence for the father at the beginning of her second year and for the
mother at 18 months proves irrefutably that the Oedipus complex
begins much earlier than at three or four years of age, as postulated by
traditional analytic theory. Rita's oedipal phase was begun but then
retreated from because of the overwhelming anxiety and guilt that the
oedipal rivalry heaped upon the preexisting anxiety and guilt of both
the paranoid and depressive positions. The witnessing of the primal
scene also stimulated excessive aggressiveness, as the little girl
believed her father was damaging her mother. Klein believed that this
case illustrates the pathogenicity of excessive aggressiveness both in
the earlier phases and in the oedipus phase. For the oedipal phase to
consolidate identifications and superego formation, the aggressive
ness of the paranoid and depressive positions must first be allayed so
that the anxiety and guilt of the Oedipus complex can be tolerated.
For the boy the original oedipal constellation is not the positive but
the negative position. Both boys and girls shift the object of desire
from the breast to the penis, but for the boy this means the adoption of
the first homosexual position (Klein, 1945). This feminine position
remains to some degree throughout life, resulting in some degree of
feminine character traits in all men. According to Klein (1945), "If the
boy can turn some of his love and libidinal desires from his mother's
breast towards his father's penis, while retaining the breast as a good
object, then his father's penis will figure in his mind as a good and
creative organ which will give him libidinal gratification" (p. 411).
Conversely, if the boy is unable to make this shift from the negative to
the positive oedipal position, the penis will become a hostile, retalia
tory, attacking organ. The crucial factors are the solidity of the good
internalized breast and the relative balance between the good and bad
internalized objects. The oral sadistic attacks on the mother's breast
Melanie Klein 103
become transferred to the penis with the result that the boy wishes to
bite off the father's penis. According to Klein, this wish constitutes the
first expression of the boy's rivalry with his father and leads to castra
tion anxiety. If the aggressiveness toward the breast is excessive, the
penis becomes a persecutory object, castration anxiety cannot be
resolved, and the Oedipus complex cannot be mastered. Further, the
boy identifies with his father's penis, so that the bad, persecutory
penis becomes his own negative masculine identification. Conversely,
if the internalization of the good breast is strong enough, the penis
will be viewed with sufficient positive affect to combat castration anx
iety. Klein believed that the boy's view of his penis, the strength of his
masculine identification, and the outcome of his Oedipus complex all
hinged ultimately on the ability of the internalized good breast to
combat the internalized bad breast. If the good object has been
irreparably damaged by the bad object in the depressive position or
split off in the paranoid position, the penis remains a persecutory
object and castration anxiety is too severe to allow for the resolution of
oedipal anxieties.
Klein's (1945, 1960) view of the boy's oedipal development is best
illustrated in her discussion of Richard, the 10-year-old patient dis
cussed earlier. Richard was a severely inhibited boy who was afraid of
other children (his refusal to go out by himself made it difficult for
him to attend school), preoccupied with his health, and given to fre
quent bouts of depression. He was precocious and gifted, and, prefer
ring adult company, he often disdained other children. The history
indicated that Richard had had a brief, unsatisfactory breast-feeding
period and was frequently ill as a child, undergoing two operations
between his third and sixth year.
The case of Richard is the only analysis Klein (1945) described fully;
the discussion of the clinical material is detailed and lengthy, but only
a fragment need be presented here to illustrate her view of the boy's
Oedipus complex. After the first interruption of the analysis, Richard
played at bumping a vampire, representing himself, into a ship named
Rodney, which represented his mother. He immediately became
defensive and rearranged the ships, including one representing his
father, in a row. This defensive arrangement, which Klein interpreted
as Richard's belief that peace and harmony could only exist in the
family if he repressed his oedipal longings, was associated with the
boy's chronic anxiety of injuring his mother. After remarks about his
mother or the analyst, he often asked, "Have I hurt your feelings?"
Richard made drawings in which red (indicating his rage, accord
ing to Klein) represented himself and blue his mother. In one drawing
104 Chapter 3
the blue parts are separated by an elongated red section that the
patient himself interpreted as a genital. Klein suggested that the object
could be a tooth and interpreted this material as "symbolizing the
danger to the loved object from the oral-sadistic impulses, the latter
the danger pertaining, as he felt, to the genital function as such
because of its penetrating nature" (p. 380). The penis represented to
Richard a dangerous object that could damage his beloved mother's
insides. Consequently, he split his mother into the idealized breast
mother and the hostile, retaliatory mother with whom he associated
genitality. Out of fear, Richard withdrew to the pregenital longing for
the breast and the idealization of the mother-infant relationship, and
consequently, was attached to his mother in an infantile way.
Klein believed that the early feeding difficulties led to overwhelm
ing oral aggressiveness and to its projected form, excessive fear of the
bad breast. Richard's paranoia was shown clearly in one session when
he frequently looked out the window and stated that two men he saw
talking were spying on him. Klein linked this persecutory anxiety to
his hypochondriacal fears, as he unconsciously feared poisoning from
his parents. This interpretation appeared to have lessened Richard's
anxieties, for the next day it appeared that his mood shifted from
depression to elation. He described how much he loved his breakfast
and how the world looked beautiful to him. Klein believed this shift to
a hypomanic defense reflected his renewed belief in his internalized
good mother.
In this session Richard proceeded to discuss two drawings from the
previous day in which his mother was represented by a "very horrid"
bird with an open beak in the colors representing himself and his
brother. His mother "now appeared as greedy and destructive. The fact
that her beak was formed by red and purple sections expressed
Richard's projection on to his mother of his own (as well as his
brother's) oral-sadistic impulses" (p. 388). It was significant to Klein that
Richard had equated this drawing with another drawing representing
himself, indicating his introjection of the devouring, retaliatory mother.
Richard assuaged his fear of the bad breast by his idealization of
the good breast, but that meant regression to the oral level. The fears
of the bad breast were transferred to his father's penis, leading to fear
of that organ and to overwhelming anxiety in the early positive oedi
pal position. His fear of his father's penis led to severe castration anxi
ety and fears of persecution, and he feared that his own penis would
injure the mother he loved. These fundamental anxieties and their
associated guilt led to repression of genitality, depression, severe
inhibitions, and fear of other children.
Melanie Klein 105
In Klein's view, Richard's transfer of the bad, retaliatory, devouring

breast onto the father's penis resulted in severe castration anxiety and
regression not only from the oedipal phase but also from the depres
sive position, since the mother was split into the genital bad mother
and the idealized good breast mother. Further, the hostile genital
breast was both a projection of Richard's sadistic wishes toward it as a
result of early frustration and constitutional deficiency, and the intro
jection of the retaliatory breast, leading to severe anxiety that pro
voked regression. Ultimately, Richard's unresolved castration anxiety
and failure to master the oedipal phase, in Klein's view, could be
traced to excessive aggressiveness in the paranoid position. She con
sidered this and similar cases to be evidence of the close connection
between persecutory, depressive, and castration anxiety and of the
impossibility of considering oedipal issues in isolation from these
early phases of development.
It is clear from this view of the Oedipus complex that Klein viewed
the pathogenicity of this developmental phase to reside in the degree
of aggressiveness toward the same sex parent which is itself ultimately
rooted in unresolved aggressiveness toward the breast. Even viewing
the primal scene, as in the case of Rita, was considered pathogenic by
Klein only because of the excessive aggressiveness it stimulated. In
Klein's view, libidinal desires are not pathogenic of themselves even
when directed to the opposite-sex parent. Klein (1945) viewed the sex
ual excitement of childhood masturbation as inevitably rivalrous with
the opposite-sex parent, leading to aggressive wishes toward that par
ent. This desire to injure provides the potential pathogenicity of child
hood masturbation. The child becomes anxious, according to Klein, not
because of the sexual excitement per se but because the rivalrous
aggressiveness threatens the parent of the opposite sex.
In the discussion of the cases of Rita and Richard, one can find
some indications of the way Klein used her theory of development
and psychopathology in treatment. Her interpretations follow her the
ory of the significance of the paranoid and depressive positions, exces
sive aggressiveness, and their influence on oedipal conflicts. However,
it is not possible on the basis of the foregoing discussion to appreciate
fully the uniqueness of Kleinian technique, and it is to her treatment
approach that we shall now turn.
TREATMENT
Klein believed herself to be an adherent of traditional psychoanalytic

principles of reliance on transference and resistance and in an absolute
106 Chapter 3
adherence to the interpretive method for resolving emotional difficul

ties (Klein, 1926,1946c, 1952d). Indeed, her strict devotion to interpre
tation was one cause of her disagreement with Anna Freud, who
believed that for children and adolescents the analyst must use non-
interpretive means to build a treatment alliance before interpretive
work can begin (A. Freud, 1927). Klein considered her primary con
tributions to psychoanalytic technique to be her emphasis on aggres
siveness in interpretations, her expansion of the psychoanalytic
method to severely disturbed adults and children, her belief in early
interpretation with children, and her advocacy of play technique for
children.
In all the case material she published, Klein emphasized in her
interpretations paranoid and depressive anxieties and their influence
on the Oedipus complex, the primal scene, and excessive aggressive
ness, as well as envy, greed, and jealousy. Oral sadism, included in
both persecutory and depressive anxiety and often expressed as fan
tasies of attacks on the mother's insides, plays a central role in her
interpretations. Klein believed these issues were present in all psycho
pathology because they are central aspects of development. Con
sequently, interpretation must be focused on aggressive wishes and
their attendant anxieties whether the central pathology is organized
around the paranoid or depressive anxieties or their manifestation in
oedipal conflicts.
One direct result of this spectrum conceptualization of psychopathol
ogy is a broadened view of analyzability. Since all forms of disturbance
are separated only by degree, the same type of treatment approach is
applicable whether the presenting symptom picture is schizophrenic,
neurotic, or in the range in between. A direct consequence of Klein's
theoretical system is the belief that psychotic cases can be analyzed by
the same methods used for neurotic cases, the issues differing only in
severity. The advocacy of strict interpretive psychoanalysis in psy
chotic and borderline cases was one of her more controversial posi
tions, but it also made her a pioneer in a trend that has been called
"the widening scope" of psychoanalysis (Bibring, 1954).
Klein's (1952d) claim for the analyzability of severe pathology is
linked to her broadened view of the transference. Klein (1952b)
believed that transference originates in the same developmental
processes that provide the initial full object relationship. Klein (1952d)
deduced from her view of the origins of mental life that this early
object relationship becomes a part of all analyses and that it forms a
major component of the transference of severely disturbed individu
als. In Klein's view, the issues of psychotic patients always manifest
Melanie Klein 107
themselves in primitive negative transferences based on these early

object relationships, which, like neurotic conflicts, can be altered by
the "good feed" of interpretations. While Klein recognized that
patients often do not make direct references to the transference for a
long time, her argument was that transference must be viewed as a
much broader phenomenon than this behavior indicates. Klein dis
puted Freud's (1912) concept of transference as a libidinal object
cathexis across the repression barrier; in contrast, she believed that the
patient transfers the "total situation" from infancy to the consulting
room. Since the transference is rooted in the earliest object relationship
and deepest layer of the unconscious, its manifestations exist from the
very beginning of an analysis, even in the patient's reporting of his
or her history. It is not just that one can see the functioning of the ego
early in the treatment; according to Klein, the patient's every pre
sentation reveals defenses against the anxiety of the transference, an
anxiety that is eventually revealed in the analysis.
For similar reasons, Klein (1946c) believed in the applicability of
psychoanalysis to children. Since childhood emotional disturbance is
rooted in the same anxieties as adult psychopathology, Klein (1926)
adopted a strict interpretive method in her work with all age groups,
even very young children. As children are often not able to verbalize
their experience, she used play to stimulate clinical material; she then
applied interpretive principles in the same way as she did with adults.
Her advocacy of this approach with children made Klein a pioneer in
child analysis.
In the treatment of children, Klein (1948c) believed strongly in early
interpretation. While she acknowledged that the material must be ade
quate to warrant transference interpretation, she also pointed out that
children provide sufficient material early in the course of treatment:
As soon as the small patient has given me some sort of insight into his
complexes— whether through his games or his drawings or phantasies,
or merely by his general behavior—I consider that interpretation can
and should begin. This does not run counter to the well-tried rule that
the analyst should wait till the transference is there before he begins
interpreting, because with children the transference takes place immedi
ately, and the analyst will often be given evidence straight away of its
positive nature. But should the child show shyness, anxiety or even only
a certain distrust, such behavior is to be read as a sign of a negative
transference, and this makes it still more imperative that interpretation
should begin as soon as possible. For interpretation reduces the
patient's negative transference by taking the negative affects involved
back to their original objects and situation [pp. 46-47].
108 Chapter 3
Although Klein's writing is replete with evidence of early interpre

tation, her examples are typically of interpretations of the Oedipus
complex, the primal scene, and the destructive impulses associated
with them and only rarely refer to the transference. For example,
Trude, age three-and-one-quarter years, in her first analytic hour
requested that flowers be removed, threw a toy man out of a cart,
wanted a man taken out of a picture book, and said that cushions had
been thrown into disorder by a dog; Klein interpreted to Trude that
she wanted to do away with her father's penis because it was playing
havoc with her mother. Her patient Peter, of the same age, in the first
hour bumped together a horse and carriage and later did the same
with two toy horses, saying, "That's not nice"; in the second session
he played again at knocking, bumping, and dangling, and Klein inter
preted that these were all symbols of his mother and father bumping
their genitals together to produce his brother. These comments are
representative of the type of interpretation Klein made to children
early in treatment, although both interpretations are genetic, not
transference, interpretations. Indeed, one of the cardinal principles of
her technique in child analysis was to reduce the intensity of affects by
continually interpreting them back to the "original situation."
In her treatment of Richard, discussed earlier, Klein (1960) did in
fact make reference to the transference in the first session. Richard
spoke of Hitler's bombs, and Klein asked if he was worried about his
mother. He replied that he worried that a tramp might break into his
mother's room and hurt her. Klein interpreted that Richard knew
Klein herself was Austrian and was also concerned about injury to
her. However, she focused her attention during the session on the
boy's mention of a tramp, which she interpreted as symbolic of the
father who Richard feared would hurt his mother with his genital.
Despite the reference to the transference, Klein's interpretive focus
was on Richard's anxiety regarding his parents' sexual intercourse as
damaging to his mother. It appears that despite her theoretical adher
ence to early transference interpretation, Klein preferred early genetic
interpretations and there is little evidence that she interpreted early
material as transference.
Klein's (1926) justification for deep, early interpretations in child
analysis was that in children the demarcation between the uncon
scious and conscious is weak, so that early interpretations can be
effective immediately. Her contention was that because children tend
to blend the unconscious with the conscious, such interpretations
strengthen the child's weak ego (although she was not clear about
how this happened). When Trude asked that flowers and other objects
Melanie Klein 109
be removed from the consulting room, she was, according to Klein,

already close to awareness of the wish to have her father's penis
removed from her mother; Klein believed that her interpretation
reduced the child's anxiety by making her wish completely conscious.
Conversely, she argued that if such material remains uninterpreted,
anxiety increases. Klein concluded that early, deep interpretations in
child analysis are not only possible but desirable. They may even
be necessary in many cases for the continuance of the treatment: if
the transference is initially negative, early interpretations are neces
sary to form an immediate positive bond with the child, thereby
reducing anxiety by tracing the aggressiveness "back to the original
objects and situation" (Klein, 1946, p. 48). If the immediate transfer
ence is positive, conditions are already present for the therapeutic
efficacy of interpretation.
These illustrations also indicate Klein's criterion of sufficient evi
dence to warrant interpretation. Although she stated that the clinical
material must justify interpretive content, these clinical vignettes
show that the child's play need not in any way be connected by the
child to the content of the interpretation. When Trude's play centered
around the notion of removing, Klein interpreted that the little girl
wanted her father's penis removed from her mother. Children's play
in itself seemed to Klein to justify its interpretation as symbolic of
early hostile fantasies and anxieties. Klein's (1952d) rationale for this
criterion is based on her broadened view of transference, as discussed
earlier, although the interpretations focused on the past rather than
the transference.
This advocacy of early in-depth interpretations without substantial
evidential basis in the associative material would seem to contradict
the time-honored analytic principle of gradually interpreting ego
defenses so as to not overwhelm the ego with anxiety and to allow a
gradual working through of defenses. Klein (1926, 1946) was very
clear in her belief that to conduct child analysis on this adult model
fails both to "establish the analytic situation" and to reduce the child's
anxiety. Resistance, she believed, increases if the child is not helped to
see the unconscious meaning of his or her play and behavior. Further,
she believed the interpretation offered must be deep because unless it
is directed to the most intense anxiety and guilt of the child's mind, it
will be ineffective and may even increase resistance. Consistent with
her stress on the pathogenicity of aggressiveness, she believed that
excessive aggressiveness and negative transference were the "deepest
strata of the mind" and that interpretation at this level loosens
the most stubborn resistances and thereby "establishes the analytic
110 Chapter 3
situation" by reducing anxiety. Klein felt that children's anxiety

required this type of interpretation, unnecessary in the adult.
Early interpretation does not resolve this deep level of anxiety, but
it "opens the door" to the unconscious and to the analysis. She recog
nized that much effort was still ahead, in helping the young patient
work through and resolve the anxiety and integrate the interpreta
tions into the ego. For example, she interpreted to Peter in the second
hour that his play with "broken men" was his wish to "kick his father
out" and his fear of his father's retaliation; seven months later these
fantasies were still being worked on in the analysis.
Klein's concept of early interpretation of material that is continually
worked on throughout the analysis can be clearly seen in her treat
ment of Richard, which is the most detailed case study she published
(Klein, 1960). As has been mentioned, Klein interpreted Richard's fear
of his father hurting his mother in sexual intercourse in the first ses
sion. In the next session Richard talked about colliding planets, and
Klein once again interpreted this as a manifestation of anxiety regard
ing sexual intercourse between the parents. Richard commented in his
response that he hated Hitler and would like to hurt him, as well as
Ribbentrop, who dared to accuse England of being the aggressor in
the war. In this session Klein added that not only was he concerned
about his father hurting his mother but that he also might be afraid his
parents were enjoying themselves, in which case he "would be jealous
and angry with them for leaving him 'lonely and deserted'" (p. 25).
She then referred to his comment about Ribbentrop, interpreting that
if he was angry and jealous of his parents, he would be the aggressor
owing to his desire to make trouble for them.
This theme—Richard's desire to injure his parents out of oedipal
rivalry rooted initially in oral sadistic desires to injure his mother—
was repeated throughout the analysis. In our discussion of the
Oedipus complex, it was pointed out that Richard's drawings, pro
duced about four weeks into the analysis, were interpreted as repre
sentations of his destructive wishes toward his mother's insides and
his fear of her retaliation, wishes that led to Richard's splitting his
maternal introject into good and bad. In the 55th session Klein inter
preted the boy's desire to injure and destroy his mother and his ana
lyst while pretending to be an innocent lamb. In the next session
Richard drew an eagle inside a coat and pulled his coat over himself
to demonstrate the eagle. Klein interpreted that he saw himself as an
eagle inside his mother's and his analyst's stomach harboring a wish
to damage their insides. Finally, as mentioned earlier, at the end of the
prematurely terminated analysis Klein interpreted Richard's anxiety
Melanie Klein 111
about her leaving as his fear that his envy and rage were so great that
he could destroy her.
Klein believed that this type of interpretation, of which only a small
illustrative sample can be given here, helped Richard to begin to intro-
ject her as a good object, a process that aided in the diminution of his
persecutory and depressive anxieties. Whatever the validity of this
claim, the same analytic issues were interpreted from the beginning to
the end of the analysis. It seems clear that in child analysis Klein
believed in interpreting the same fundamental issues throughout the
analysis rather than allowing a gradual unfolding of material.
As can be seen from these clinical examples, the content emphasis
in Kleinian child analysis is no different from the interpretive focus
in adult analysis. Klein believed that the child's unconscious aggres
sive feelings toward the parents, including all doubts and criticisms
of them were central in all childhood conflicts. She also felt that the
making conscious of sexual fantasies is crucial to the reduction of anx
iety; however, she believed that the pathogenic component of these
sexual fantasies lies in their aggressive nature, which the child's mind
perceives as a threat to the parents.
Klein (1945) believed that oedipal conflicts were central to pathol
ogy but could not be separated from persecutory anxieties. In the case
of Rita, Klein's oedipal interpretations focused on the child's hostility
toward her parents in sexual intercourse. For example, Rita played
that she was traveling with her teddybear to a "good woman" who
was to give her a treat, but on the way she got rid of the engine driver
who kept returning to threaten her and they battled over her teddy-
bear. Klein interpreted that the bear represented her father's penis,
which she had stolen in order to take his place with her mother. To
this oedipal interpretation, Klein added that the fantasied penis rob
bery was an effort to repair the fantasied damage done to the mother's
body by her oral sadistic attacks. Recall that Rita had witnessed the
primal scene. Klein interpreted to her that when she observed coitus
she wanted to do with her father what her mother was doing, that she
sadistically wished to injure her mother out of jealousy. Klein inter
preted the teddy bear play as evidence of Rita's failure to overcome
her sadistic wishes toward her mother.
Similarly, Rita had a bedtime ritual of being tightly tucked in with
her doll to prevent a mouse or "butzen" (her word) from coming in
through the window to bite off her "butzen." Again, Klein interpreted
the oedipal level: Rita feared that her father's penis would bite off her
imaginary penis just as she desired to castrate him. However, Klein
added that the fear of entry through the window also represented the
112 Chapter 3
fear of her mother's retaliatory attack upon her for her sadistic wishes
to attack her mother's body. The phobias and nighttime terrors were
all interpreted as Rita's fear of maternal retaliation, and her wish for a
penis as a wish to repair her mother's damaged body. Klein also inter
preted Rita's inability to separate from her mother as evidence of her
continual need for reassurance that she had not damaged her mother
with her sadistic wishes. Klein considered her difficulty accepting the
bottle and then her resistance to yielding it for food as symptomatic of
this same anxiety.
One can see from these examples that Klein's interpretive focus for
the solution of Rita's multiple symptom picture was on the little girl's
unconscious hostility toward the parents and the resulting depressive
and persecutory anxieties. Klein's view of Rita's oedipal conflicts, as
we have seen, was that they were pathogenic because of her unre
solved oral sadistic wishes toward her mother's body, which resulted
in the persecutory anxiety of retaliation and the depressive anxiety of
having done fantasied injury to her mother's body. It was this inter
pretive focus on the anxiety of doing damage to the mother and the
resulting fears of retaliation and loss that Klein felt was mutative in
her treatment of Rita.
This emphasis on aggressive wishes was also seen in the case of
Ruth, a four-and-one-half year-old who refused to be alone with Klein
(1946), and thus treatment had to be conducted with Ruth's older sis
ter in the consulting room. The child had an overly strong attachment
to her mother and some, but not all, other women. She was timid, had
great difficulty making friends, and suffered from severe anxiety.
Once when the sister was ill, Ruth had to enter the consulting room by
herself, which led to a severe anxiety attack; meanwhile Klein played
at feeding dolls, as Ruth had done in the previous session. When the
analyst put a wet sponge near one doll, Ruth screamed that the big
sponge was only for grown-ups and must not be given to the doll. As
in the previous session, the material related to envy of the mother;
Klein now interpreted that Ruth "envied and hated her mother
because the latter had incorporated her father's penis during coitus,
and .. . wanted to steal his penis and the children out of her mother's
inside and kill her mother" (p. 56). Klein went on to explain to Ruth
that her fears were due to her anxiety of having killed her mother with
her rage. Klein reports that after this interpretation Ruth's anxiety
dissipated considerably in the session and that after a few more ses
sions she had very little anxiety entering the consulting room. Klein
analyzed Ruth's anxiety attacks as a repetition of pavor nocturnus,
which she suffered at age two when her mother became pregnant
Melanie Klein 113
and she wished to steal the new baby and kill it. The guilt from this
wish resulted in her over-attachment to her mother, with Ruth
needing to be constantly in her mother's presence to reassure her
self that she had not killed her mother. When she went to sleep,
Ruth feared she would never see her mother again. In this case,
as in the case of Rita, the source of the pathology, according to
Klein, is to be found in the sadistic wishes to attack the mother's
body, resulting in both the persecutory anxiety of retaliation and
the depressive anxiety of guilt and fear of loss. The case of Ruth
also demonstrates the increasingly central role envy came to play
in Klein's understanding of the pathogenicity of aggressiveness.
Ruth did not simply hate her mother but envied her, and Klein
believed that envy was the root of Ruth's wish to damage her
mother and her mother's body. For Ruth, unlike Rita, the full force
of the anxiety was felt toward the analyst and was the basis for an
intense negative transference.
In accordance with her view of the pathogenicity of aggressive
and envious object relationships, Klein considered the negative
transference to be a critical component of all analytic treatment,
since the patient projects the early pathogenic hostile object relation
ship onto the analyst. Since envy is inevitably an aspect of the trans
ference, the analyst's interpretations, especially if they are good and
potentially helpful, are envied. In Klein's (1957) view, a good inter
pretation symbolizes a good feed. Therefore, to accept a helpful
interpretation is to acknowledge that the analyst has good supplies
that the patient lacks; envy follows, along with its desire to devour.
To defend against envy of the analytic "good breast," the patient
may reject the interpretation. Klein was well aware of the possible
dangers of this view: if misused, it allows the analyst a way to blame
the patient's rejection of interpretations on the patient's envy.
However, she did not view all patient rejections of analytic interpre
tations as envy; she was referring only to interpretations that have
been helpful. Recall that the infant is envious of the good breast for
having and of the bad breast for withholding. The same phenomena
become an inevitable component of the negative transference, fur
ther intensifying it. The powerful force of envy is, in Klein's view,
the primary source of patient resistance. By viewing resistance in
this light, Klein shifted an analytic concept from its original libidinal
meaning to an aggressive phenomenon. She felt that resistance to the
awareness of libidinal drives, such as oedipal longings, did not rep
resent as powerful an obstacle to treatment progress as did the
patient's envy of the analyst.
114 Chapter 3
Indeed, Klein viewed the resolution of envy as a major aspect of the

analytic process, believing that all patients envy the analyst who rep
resents the good breast and consequently have some degree of hostil
ity to the analyst for being helpful. The analyst's task is to recognize
and interpret this resistance, which relieves anxiety immediately but
which itself eventually results in envy. Analysis tends to assume a
cycle in which "good feeding" interpretations that relieve anxiety are
followed by envy of the analyst/good breast for having this "good
food" to offer, which leads to the need to spoil and increases resis
tance, which the analyst then interprets, thus offering more "good
food." The good experiences of accurate interpretations will in each
instance relieve anxiety and lead to the introjection of the good object.
This side of the process must ultimately win out over envy if the
analysis is to be successful. When the good object is successfully intro
jected, the patient is able to feel gratitude toward the analyst rather
than simply envy. Gratitude reflects a strengthened introjection of the
good object and allows the integration process to unfold. Just as in
infancy the crucial factor in overcoming envy and aggressive wishes
toward the breast is the introjection of the good breast, so too in treat
ment the good feeds, represented by interpretations have the potential
to overcome envy, greed, and aggressiveness. Klein did not believe in
the provision of a good therapeutic experience, as Guntrip (1969) did.
Her theory of treatment was orthodox in its strict adherence to inter
pretation as the only appropriate analytic intervention. In Klein's
view, the accurate interpretation represents the good feed that allows
the patient to introject the analyst as a good object and eventually
leads to ego integration.
If the patient's envy is excessive, resistance will be enacted in some
way to defeat the help offered by the analyst. Every step of progress in
the analysis is followed by an act designed to destroy it. Klein
believed that excessive envy explained the need of many charactero-
logically disturbed patients to devalue the therapeutic process, espe
cially when it has been helpful. She was referring to patients who after
helpful interpretation feel some degree of relief from anxiety and new
hope and then attack the very interpretation that brought them relief.
Klein (1957) described the ensuing therapeutic process with such a
patient this way:
His criticism may attach itself to minor points; the interpretation should
have been given earlier; it was too long, and has disturbed the patient's
associations; or it was too short, and this implies that he has not been
sufficiently understood. The envious patient grudges the analyst the
success of his work; and if he feels that the analyst and the help he is
Melanie Klein 115
giving have become spoilt and devalued by his envious criticism, he

cannot introject him sufficiently as a good object nor accept his interpre
tations with real conviction and assimilate them . . . . The envious
patient may also feel, because of guilt about devaluing the help given,
that he is unworthy to benefit by analysis [p. 184].
The negative therapeutic reaction, in Klein's view, is not simply a

matter of the patient feeling guilty; the patient envies the analyst and
desires to scoop out the good food from the analyst and spoil it.
Receiving anything good from the analyst is associated with the fan-
tasied injury to him or her. To protect the good object, the patient does
not tolerate being given to, and must sabotage any offer of help.
The treatment of the negative therapeutic reaction is illustrated by
the analysis of a female patient who was convinced that her babyhood
and early feeding had been unsatisfactory (Klein, 1957). One day the
patient phoned to cancel two consecutive sessions because of shoulder
pain. When she did come, she complained extensively of pain and of
others' lack of interest in her. She felt a need for someone to cover her
shoulder, make her feel warm, and then go away. It occurred to her
that this must be how she felt as a baby when she wanted to be cared
for and no one came. The patient reported a dream in which no one
served her in a restaurant but a determined woman in front of her
took two or three cakes and went away; the patient then took two or
three cakes herself. According to Klein (1957), the patient's
grievance about the missed analytic sessions related to the unsatisfactory

feeds and unhappiness in babyhood. The two cakes out of the "two or
three" stood for the breast which she felt she had been twice deprived of
by missing analytic sessions . . . . The fact that the woman was "deter
mined" and that the patient followed her example in taking the cakes
pointed both at her identification with the analyst and at projection of her
own greed onto her . . . . The analyst who went away with the two or
three petit fours stood not only for the breast which was withheld, but also
for the breast which was going to feed itself [p. 205].
Klein pointed out to the patient that her frustration had "turned to
envy," as the mother/analyst was suspected of enjoying herself while
the patient was missing the analytic sessions. Klein's interpretation of
the dream focused on the connection between the missed analytic
sessions and the unsatisfactory breast experience, both of which made
the patient feel envious and resentful. According to Klein, the patient
had felt her mother "to be selfish and mean, feeding and loving
herself, rather than her baby" (p. 205).
116 Chapter 3
Klein reported that the interpretation of this dream resulted in a

dramatic emotional shift in the patient, who felt a happiness and grati
tude that made her feel as if she had had a perfectly satisfactory feed.
The patient then related that her early feeding may have been better
than she had thought, and for the first time she recognized her envy of
the analyst and her desire to spoil both the analyst and the analysis.
This session began a process of working through the shift from envy
and the desire to spoil to gratitude and enjoyment. The dream interpre
tation illustrates Klein's technical principles in analyzing resistance and
negative therapeutic reactions. The patient's envy of and hostility
toward the analyst were interpreted and related to the early breast
experience, and the interpretations themselves counteracted the aggres
siveness by providing good feeds that resulted in the introjection of the
"good analytic breast," leading to gratitude and enjoyment.
The interpretation of the negative transference and envy are no less
significant in the analysis of depressive position dynamics of ambiva
lence toward whole objects (Klein, 1937,1957). The primary difference
between the analysis of dynamics in the depressive and paranoid
positions is the focus on anxiety of damage to the object in the former
position as opposed to fear of injury to the ego in the latter. In both
cases the analytic process involves the gradual integration of love
and aggression through the continued interpretation of the conflict
between them.
Klein's approach to the treatment of fixation at the depressive
position is illustrated in her discussion of a male patient who had not
been fully aware of his destructive desires and guilt about them
when he reported a dream about fishing (Klein, 1957). In the dream,
the patient decided to put his fish in a basket to die rather than eat it.
He was carrying the fish in a laundry basket when it turned into a
beautiful baby with something green about it; he noticed that the
baby's intestines were protruding because of the hook it had swal
lowed. The green color was associated with Klein's green books, and
the patient stated that the fish stood for one of Klein's own books. In
addition, Klein (1957) states, "the fish was not only my work and my
baby but also stood for myself. My swallowing the hook, which
meant having swallowed the bait, expressed his feeling that I had
thought better of him than he deserved and not recognized that there
were also very destructive parts of his self operative in relation to
me" (p. 211). After the interpretation, the patient became deeply
depressed as he was horrified to realize the depths of his destructive
desires and envy. Klein interpreted this reaction as a response to a
step toward integration.
Melanie Klein 117
The next night the patient dreamed of a pike, to which he associ

ated whales and sharks although the pike was old and worn rather
than dangerous. A suckerfish was on it, and the patient immediately
pointed out that the suckerfish protects itself this way. Klein's (1957)
analysis of the dream is as follows:
The patient recognized that this explanation was a defence against the
feeling that he was the suckerfish and I was the old and worn-out pike
and was in that state because I had been so badly treated in the dream
of the previous night, and because he felt I had been sucked dry by
him. This had made me not only into an injured but also into a danger
ous object. In other words, persecutory as well as depressive anxiety
had come to the fore; the pike associated to whales and sharks showed
the persecutory aspects, whereas its old and worn-out appearance
expressed the patient's sense of guilt about the harm he felt he had
been doing and was doing to me [p. 212].
The dream illustrates not only the patient's fear of having injured the
analyst with his destructiveness and envy, but also a regressive reac
tion from depressive to persecutory anxiety, illustrating the mixture of
depressive and persecutory anxiety often found in cases of depres
sion. After the interpretation of the dream, the patient underwent an
intense depression for several weeks during which his urge for repa
ration intensified. When he did emerge from his depression, the
patient felt that his knowledge of himself increased to such an extent
that he would never again see himself as he had in the past and that
his tolerance of others had also improved. This step toward integra
tion could not be sustained initially; the patient continually lapsed
into depressive states, which were interpreted as guilt over injuring
the analyst.
The analysis of the patient's guilt led to a realization regarding his
need for reparation: "An overstrong identification with the object
harmed in phantasy—originally the mother—had impaired his capac
ity for full enjoyment and thereby to some extent impoverished his
life" (p. 213). Klein believed that the patient's early breast-feeding had
not been completely satisfying because of his fear of exhausting and
depriving the breast. Klein concluded that he had experienced guilt
over envy of the breast too early, leading to persecutory anxiety,
which explained his regressive response to the awareness of depres
sive anxiety. The interpretations of unconscious hostility, envy,
and guilt focused on the analyst allowed the gradual integration of
love and aggressiveness, resulting in an increase of gratitude and
enjoyment and the eventual lifting of the depression.
118 Chapter 3
This case demonstrates Klein's (1950) view of treatment goals and

her criterion for a good psychoanalytic outcome: resolution of perse
cutory and depressive anxiety. Nowhere does Klein give any indica
tion that treatment goals differ by diagnostic category; in her schema
all patients seem to need to achieve the same goals. Differences among
patients have to do with severity and the developmental phase of anx
iety, not its content. Patients resolve their symptoms to the extent that
persecutory and depressive anxieties are mastered in treatment, as
indicated by the degree of introjection of the good object. Persecutory
anxiety diminishes as the internalization of the good object mitigates
the sense of danger from without, and depressive anxiety is assuaged
as the fear of danger to the good object is reduced. As we have seen,
for Klein these critical psychic shifts occur in response to good inter
pretations that reduce anxiety and foster the establishment of the
good object within. The strengthening of the good object allows it to
survive destructive impulses and contact with the bad object without
suffering severe damage, thereby reducing the need for splitting and
fostering whole object integration. Ultimately, Klein viewed the out
come of successful analytic treatment as the integration of the ego and
whole objects.
CRITIQUE
Klein's views have been controversial from their inception and have
stimulated so much intense criticism and debate within the psychoan
alytic community that an appraisal of their strong and weak points is
in order. After Klein developed the concept of the depressive position,
her work came under intense attack from many traditional analysts.
Glover (1945), the leader of the anti-Kleinian group, caustically criti
cized the "new metapsychology." Glover's hostile tone represented
the atmosphere within the British Psychoanalytic Society at the time,
where many traditional analysts felt Klein's views were highly specu
lative, even fantastical, and represented a departure from the founda
tion of psychoanalysis (for example, Waelder, 1937; A. Freud, 1927).
These analysts were absolute in their rejection of Klein's views, result
ing in a split between the Kleinian and anti-Kleinian forces within the
British Psychoanalytic Society (Segal, 1980). It is only in recent years,
as interest in object relations theories has intensified, that some
Kleinian concepts have been given serious consideration beyond her
group of devoted followers. Her theoretical system has been attacked
on virtually every point, but the following five general criticisms are
most compelling.
Melanie Klein 119
First, Klein has been widely criticized for "adultomorphism," that

is, attributing to infants the mental processes of adults. Waelder (1937)
was the first to set forth a thorough critique of the theory that infants
have complex fantasy lives that include desires to injure, fear of retali
ation, good and bad objects, and oedipal conflicts. Even if such fan
tasies are seen in the analysis of children as young as the third year,
say anti-Kleinians, the assumption of their existence in the first year is
unwarranted. Bibring (1947) pointed out that Kleinian theory assumes
complex fantasied relationships among objects long before perception
has developed to the point that such fantasies would be possible. This
criticism by the anti-Kleinians seems well justified by research on
infancy, which shows clearly that infants do not have the cognitive
capacity to have the complex fantasy life Klein attributes to them
(Lichtenberg, 1983, Stern, 1985). As a corollary to this criticism, Klein's
belief that infants have knowledge of sexual intercourse is without
basis; indeed, such knowledge is not possible given the cognitive
equipment of infancy. Even children who do in fact witness the primal
scene in the first few months of life have such a limited cognitive
capability that it is impossible for them to fantasize that the father's
penis is inside the mother and to have the wish to do the same them
selves. Besides, even if one could argue that such cognitive capability
exists in young infants, the evidence that they in fact have such
fantasies is lacking.
The second major criticism of Klein's developmental theory is that
she confuses psychopathological constellations with normal develop
mental process. She attributes the pathological reactions of paranoia,
depression, mania, and obsessions to the normal infant, yet the only
data she offers for this are her analyses of pathological children. Even
if one agrees with her formulations of pathological mechanisms in
these children, the vast research on infancy offers no evidence for
the existence of these pathological constellations in normal infants
(Stern, 1985). In Klein's defense, she was aware of this criticism and
attempted to clarify her position by stating that the infant is not psy
chotic but has "psychotic-like" anxieties (Klein, 1946). While this clari
fication mitigates Klein's position somewhat, it does not resolve the
problem; evidence that the normal infant has "psychotic-like" anxieties
and complex defenses is unconvincing. From the fact that screaming
infants are comforted, Klein (1952a) inferred that they feared attack
from a persecutory object and were calmed when comforting allowed
for the reestablishment of the good object. This "evidence" is insuffi
cient for even the somewhat softened conclusion that children suffer
"psychotic-like" anxieties. The fact that children scream does not
120 Chapter 3
warrant the conclusion that they feel persecuted, and their return to
comfort does not imply the "re-establishment of the good object."
This critique points to difficulties inherent in Klein's concept of the
aggressive drive, the third general area of difficulty with her system.
Klein's use of the concept of aggression is far too loose: she inter
preted infant distress as aggression, which she then equated with
hatred, the desire to destroy, and sadism. Not only does distress not
imply aggression, but even a legitimate aggressive response is not
inherently hateful. As we will see in the discussion of Kernberg in
chapter 5, joyful assertiveness is aggressive but not hateful. Sadistic
wishes are pathological distortions of aggression and should not be
equated with it. Klein did not make crucial distinctions between the
infant's distress, joyful aggressive expressions, healthy assertiveness,
and hateful forms of aggressiveness. Moreover, there is no basis
for the assumption that aggression is a drive (as will be discussed fur
ther in chapter 5). Aggressive responses are not motivated by biologi
cal pressure for gratification, as are such drives as sex, hunger, and
thirst (Scott, 1958). Klein based her developmental theory on a faulty,
confused concept of aggression.
The fourth area of difficulty with Klein's theory involves the incon
sistencies in her attempted distinction between the paranoid and
depressive positions. Although many analysts have found the distinc
tion to be generally useful, Klein's contention that destructive, sadistic
desires exist in the paranoid position conflicts with her view that the
desire to injure the object first appears in the depressive position. It
may appear that Klein's view could be defended by reasoning that
aggression in the depressive position is the first experience of intent to
injure the loved object whereas the destructive desires of the paranoid
position are directed toward the split-off bad object. However, the
purported existence of envy in the paranoid position weakens this line
of reasoning: envy is directed against the good object, and Klein's the
ory must somehow reconcile the envious desire in the paranoid posi
tion to damage the good object with the contention that the depressive
position is defined by the first desire to injure the loved object.
The fifth major category of criticism of Klein's work is that her
interpretive style involves unwarranted, even wild, inferential leaps
that contradict sound analytic principles and lack evidential basis in
the clinical material (Zetzel, 1956, 1964). Critics point to Klein's char
acteristic tendency to interpret children's play material as symbolic of
the primal scene and other primitive fantasies without waiting for
confirming associative material. The criticism applies also to adult
analysis because, although Klein did not specifically advocate early
Melanie Klein 121
interpretation in adult analysis, she used the same principles to

deduce transference, the primal scene, and paranoid and depressive
anxieties from symbolic material, deductions that sometimes lacked
confirming associative data. Critics have argued that Klein ignored
patients' needs for defenses and therefore the need to work through
them gradually in order for the ego to be ready for the awareness of
the unconscious material it had so staunchly resisted (Kernberg, 1972).
Zetzel (1964) pointed out that interpretation implies a therapeutic
alliance with part of the patient's ego, an alliance that allows the inter
pretive process to occur. Kernberg (1972) summarized these criticisms
as the "ego-psychological critique" of Klein's work, because he felt
Klein ignored one of the basic postulates of ego psychology, namely,
that the ego and its defenses must be central to analysis before id
interpretations can have a mutative effect.
As a corollary to the "ego-psychological critique" of Kleinian inter
pretive principles, critics (Kernberg, 1972) have suggested that Klein's
early, deep interpretive style resulted in a lack of analytic process—
the same material is addressed and interpreted at the same level from
beginning to end. Klein herself acknowledged this persistence of inter
pretive themes in the case of Peter, but she apparently did not con
sider the lack of analytic process a problem. Geleerd (1963) found fault
with Klein's analysis of Richard because the analysis never deepened;
the same content was interpreted at the same level from first session
to last. Geleerd pointed out that the result was a lack of discernible
analytic progress.
Despite the serious flaws in Klein's views, she was able to explain a
great deal of psychopathology by differentiating persecutory and
depressive anxieties, demonstrating their role in a wide variety of psy-
chopathological conditions, and illuminating the role of primitive
defenses in neurosis and character pathology. Both Kleinians and non-
Kleinian clinicians have found her notions of splitting, projective and
introjective cycles, and the mechanisms of denial, omnipotence, and
especially, projective identification to be of inestimable value in the
treatment of severe character pathology. (Kernberg, 1975; Giovacchini,
1979; Ogden, 1982; Grotstein, 1986; Hughes, 1987). Projective identifi
cation has become widely used as a tool for the understanding of
character pathology. In addition, a variety of symptoms and patholog
ical patterns—such as food addictions, excessive object attachment,
depression, and hypochondriasis—can be explained, at least in part, by
Klein's understanding of depressive position dynamics, and the differ
entiation of these dynamics from paranoid anxiety provides the clini
cian with an effective tool for the separation of levels of pathology.
122 Chapter 3
Further, Klein's view that ego and superego development are products
of the internalization of object relations has been found useful by
object relations theorists. We have already seen, in chapter 2, that
Fairbairn and Guntrip used her concept of internal objects to develop
their own conceptualizations of development and psychopathology
and employed her concept of the depressive position to differentiate
categories of psychopathology.
None of the aforementioned criticisms of Klein's views necessarily
contradicts the object relations foundation of her view of develop
ment and psychopathology. One can agree with these major criti
cisms and continue to adhere to Klein's position that the crucial
factor in early development is the initial object relationship and the
relative balance of good and bad objects in this relationship. One
need only change the Kleinian developmental timetable to the later
phases of childhood, when the child has the cognitive and perceptual
apparatus to perceive objects and their interrelationships, to escape
the difficulties of adultomorphism while preserving Klein's object
relations view of development and psychopathology. In this way, the
substance of Klein's theory, that paranoid and depressive anxieties
characterize the conflicts of early stages of development and that
excessive conflict in these stages predisposes the ego to a pathological
outcome later in development, is largely unaffected by the aforemen
tioned criticisms. Klein's emphasis on object relations is consistent
with the traditional psychoanalytic conception that conflicts and anx
iety in each developmental phase leave the ego prone to some degree
of fixation or regression that may result in pathology, but that these
are also elements of the normal developmental processes does not
necessarily include pathological elements. Finally, Klein's aggressive
interpretive style has no inherent connection to her emphasis on
object relations. Her view of development as consisting of good and
bad internal object relations can as easily be applied clinically accord
ing to the technical precepts of developing an analytic relationship
and sensitive interpretive timing. Indeed, Segal (1981) and Rosenfeld
(1987) adapted Klein's theoretical concepts to a more cautiously inter
pretive analytic model. These and other modifications to Klein's basic
theoretical postulates have been made by Klein's followers.
THE KLEINIANS
Despite the intense and frequent criticism with which her views were
met by many analysts, Klein attracted a devoted group of followers
who adopted most of her fundamental ideas. Her devotees modified
Melanie Klein 123
and expanded many of her ideas, but they adhered to her basic con
cepts—the paranoid-schizoid and depressive positions,and the impor
tance of the projective-introjective cycles, splitting in development
and pathology—and a Kleinian school was born within the psychoan
alytic movement (although it did take her theory in directions she
probably did not anticipate). The most striking additions to Klein's
thought among her followers fall into five general categories: (1) the
expansion of the concept of projective identification; (2) the differenti
ation of clinical syndromes and their specific mechanisms and treat
ment approaches, based on Kleinian concepts; (3) the application of
Kleinian technique to psychotic states; (4) the inclusion of noninter-
pretive techniques in treatment; and (5) the evolution of the treatment
model to an emphasis on countertransference.
Projective identification has been used by the Kleinian school to
reconceptualize the psychoanalytic theory of early development and
psychopathology, as well as its principles of technique. The basic
modifications in the concept have been its extension to an interper
sonal process and elaboration beyond its defensive function (Segal,
1981; Rosenfeld, 1987). The predominant view among the Kleinians
has been that the object must in fact experience the projection in order
for an affect or unwanted part of the self to be successfully projected
into it. The self gets rid of the distressing feeling by giving it to the
other to feel. Whereas for Klein projective identification was a fantasy,
it is for her followers an interpersonal process in which the object has
the experience the self is not able to have. Rosenfeld (1987), for exam
ple, points out that projective identification is used by both infants
and adults not only for defensive purposes but also for communica
tion to let others experience one's feelings directly when words may
not be direct or forceful enough. As we shall see in chapter 7, the con
cept of the patient communicating by having the analyst feel aspects
of the self has become an important component of the interpersonal
school of psychoanalysis.
For Bion (1959a, 1962), projective identification is a crucial compo
nent of the early child-mother interaction. The infant attempts to rid
itself of distress by projecting it into the mother. In Bion's view the
primary role of the mother in early infancy is to be a "container" for
the frustration and pain the child's infantile ego is too fragile to con
tain. The mother not only contains the tension but also gives it back to
the infant in a tolerable form. The mother's ability to soothe the infant
in distress is dependent on her capacity to absorb the infant's tension
and to allow the infant to internalize her as an object capable of toler
ating the original anxiety (Segal, 1981). In this way the infant becomes
124 Chapter 3
capable of managing frustration and anxiety, thus acquiring a primary

foundation of mental stability. That is to say, the mother's ability to
allow the infant's projective identifications plays a crucial role in its
eventual mental health.
The corollary of this view of early mothering is that if the mother is
unable to contain the infant's distress, the child is left with over
whelming anxiety and is forced to deny reality. In extreme cases the
denial leads to psychosis. According to Bion's (1961) particular way
of formulating psychotic thought process, which is based on Klein's
theory of the early projective and introjective processes and his own
theory, thinking originates in a mating of a preconception and a frus
tration. If the infant expects a breast and none is forthcoming and if a
minimal capacity for frustration exists, the thought "no breast" will be
created to bridge the gulf between the want and the satisfying action.
Bion (1962) calls this process "learning from experience." The devel
opment of thought, in this view, both depends on the capacity for
frustration tolerance and increases it. If the mother cannot fulfill her
function as container, the infant does not have enough frustration tol
erance for thought development to occur (Bion, 1959a). Instead, the
infant evades frustration by treating thoughts as bad objects that must
be evacuated and by attacking the links between thoughts in an effort
to deny reality. The result is a fragmenting of the thought process and
eventual psychosis (Bion, 1957). Further, the infant is forced into a
desperately excessive use of both projective identification and split
ting in an effort to rid itself of frustrating experience; the outcome is
forceful effort to enter the object, which results in psychotic delusions.
In Bion's (1957) view, even if the infant's attempts at projective
identification are successful the object does not completely tolerate the
anxiety, and when it is reintrojected it becomes attacking and starves
the personality of all good qualities. If the infant's intolerance for frus
tration is too great—yet not enough for evasion—the outcome will be
the use of omnipotence to avoid reality. In this case, omniscience
becomes a substitute for learning from experience, and reality is
denied, but thought does not become fragmented and the resulting
personality organization is borderline rather than psychotic.
Rosenfeld (1965, 1983) adds a further component to the Kleinian
view of psychosis: he points out that the psychotic not only needs des
perately to project unwanted parts of the self into others but also
introjects others just as forcefully into the self. Both processes result in
the blurring of self-object boundaries, which the psychotic needs in
order to defend against the awareness of separateness. Rosenfeld
agreed with Bion that the mother must be a container for the infant's
Melanie Klein 125
projective identifications, but he believed that the result of failed con

tainment is excessive envy and aggressiveness and an intense need to
defend against them. The blurring of self-object boundaries, accord
ing to Rosenfeld, is the psychotic defense against intolerable envy,
aggressiveness, and the awareness of dependence.
Rosenfeld (1978) agreed with Bion that borderline psychopathology
is attributable to a failure by the mother to introject the infant's projec
tions. However, Rosenfeld put more emphasis on the resulting
increase in envy and aggressiveness, which, he believed, disrupts the
normal splitting process and thus causes prolonged states of confu
sion in which love and hate are undifferentiated. When strong affects
are evoked, the child resorts to pathological splitting of ego and object
into fragmented bits, a process leading to loss of the reality sense. In
therapy, these patients, unlike psychotics, are able to maintain a sense
of reality outside the transference and to have other relationships that
do not evoke strong affects. However, the reality sense is always vul
nerable to potential disruption by emotional contact or eruption of
affect. Rosenfeld attributed the chaos of the borderline patient to the
ease with which affect disrupts the reality sense.
Envy is defended against by attacking the dependent, libidinal self
in an effort to destroy links to objects. The result is a primitive, attack
ing superego and idealization of the destructive, omnipotent self that
disdains contact with objects. The child resorts to a fixated omnipo
tence in which reality is denied, yet the thought process is not totally
disturbed, as it is in the psychotic solution of attacking thought itself.
This formulation fits Bion's concept of the "in-between" childhood
state in which the mother fails the infant's need for projective identifi
cation enough so that there is insufficient ability to tolerate frustration
but not so much frustration that total evasion of it is necessary.
Rosenfeld adds that since the failure of projective identification arrests
development before the normal superego can develop, the attacks on
the dependent part of the self substitute for normal superego develop
ment. Moreover, this primitive superego is projected onto others,
leading to the persecutory fear that others are constantly critical and
attacking.
Klein's followers have also delineated narcissistic pathology in
a more precise way than Klein did herself. Klein (1946c) referred to
narcissism in the context of the narcissistic object ties in schizoid
mechanisms. Her followers used her concepts to formulate a view of
narcissistic pathology as a syndrome in itself. According to both Segal
(1983) and Rosenfeld (1971,1978), the narcissistic character structure is
a defense against envy and dependence.
126 Chapter 3
Rosenfeld viewed narcissistic pathology, like borderline conditions,

as a product of such excessive reliance on projection and projective
identification that the self-object distinction becomes blurred. The dif
ference between narcissism and borderline pathology is in the narcis
sistic patient's ability to utilize splitting effectively, that is, confusion
is not a dominant component of the narcissistic personality. Rosenfeld
saw the fusion of self and object in narcissistic states as a defense
against the awareness of separateness, which stimulates envy, depen
dence, and frustration. The strength of this fusion is a function of the
intensity of the destructive desire and envy that result from early frus
tration, failed projective identification in infancy, and constitutional
disposition.
Rosenfeld (1971) differentiated libidinal and destructive narcissism:
the former is the idealization of the self by the omnipotent introjection
and projective identification of good objects while the latter refers to
the idealization of the omnipotent destructive parts of the self that
attack the libidinal self and libidinal object relationships that seek
dependence ties. In narcissistic pathology the sense of humiliation
regarding needs is so deep that the destructive narcissistic self wishes
to dominate the entire personality and destroy the libidinal self in
order to eliminate all possibility of dependence ties. Rosenfeld identi
fied the dependent self as the sane part of the personality and felt that
destructive narcissism is pathological. The clinician cannot usually
discern the difference between the two types of narcissistic structure
until the self-idealization is threatened by self-other awareness. At
that point the eruption of envy and destructive desires leads to an
attack by the destructive narcissistic self on the libidinal self, and the
difference becomes clear as destructive omnipotence threatens to take
over the entire personality; the patient now displays a superior,
hostile posture, devaluing others and denying need for all objects.
Segal's (1983) formulation agrees with Rosenfeld's in viewing
extended narcissism as a defense of withdrawal against envy and
dependence. The difference in formulations is that in Segal's there is
no role for libidinal narcissism. All prolonged narcissism, according to
Segal, is based on excess aggression. The patient maintains a hostile,
superior defensive structure to defend against envy and destructive
desires; thus, envy is the crux of narcissistic pathology.
In both Segal's and Rosenfeld's conceptualizations of narcissism,
Klein's concept of the pathogenicity of excessive envy is used to
understand fixation in pathological narcissism, a syndrome not dis
cussed by Klein herself. Narcissistic pathology, the borderline syn
drome, and psychotic states are all syndromes that Klein did not
Melanie Klein 127
delineate specifically, but that have been formulated by her followers

in accordance with her theoretical concepts.
These views of Klein's followers have clear implications for treat
ment, the most controversial of which is the application of psychoana
lytic treatment to psychotic conditions. While Klein treated only one
psychotic patient, a five-year-old boy, her devotees carried out more
systematically her view that such conditions are analytically accessible
(Segal, 1980). Both Bion's and Rosenfeld's formulations of psychosis
rely on the concepts of excessive aggressiveness and envy in the
paranoid-schizoid position and the overreliance on projective identifi
cation. The clinical implication is that the analysis of envy, aggressive
ness, and projective identification has the potential to resolve
psychotic conditions. This view is the theoretical foundation for the
treatment of psychosis among Kleinian analysts, each of whom
emphasizes a different aspect of the process.
Rosenfeld (1987) emphasized the implications of the psychotic
patient's reliance on projective identification for the transference, pos
tulating that as soon as the patient enters the analyst's office he or she
attempts to communicate by making the analyst feel unwanted parts
of himself or herself. This is a mode of communication that is central
to the psychotic patient, since verbal thought for such a person has not
developed the meaning it has for the normal person. The analyst's
first task is to understand the patient's nonverbal communications by
accepting and understanding the patient's projections into him or her,
since the mother's failure to do so in childhood is the presumed
source of the psychosis. This therapeutic use of countertransference is
a direct application of Bion's (1959, 1962) view that the mother must
contain the infant's frustration for the ego development to proceed. The
attitude of empathy and acceptance of the patient's unconscious com
munication by projective identification is considered to be as crucial
to the success of the treatment as the accuracy of interpretations.
In Rosenfeld's view, psychotic patients also project their own use of
projective identification onto the analyst and therefore fear the analyst
will attempt to take over their personality: the belief that the analyst is
forcefully intruding his or her feelings into them blurs self-object
boundaries and leads to the formation of a psychotic transference. The
resolution of the transference psychosis is the crux of the treatment.
Before it can be analyzed, the analyst's task is to maintain the bound
aries by continually sorting through the patient's projections into him
or her to differentiate the patient's projections from his or her own
feelings. A crucial component of the treatment is the analyst's holding
the projective identifications and verbalizing them to himself or
128 Chapter 3
herself before an interpretation is made. In fact, Rosenfeld warned

against interpreting too quickly, since premature interpretation will be
felt by the patient as an expulsion from the analyst, and therefore as a
repeat of early trauma. (Allowing for the unfolding of the transference
constellation before interpretive work begins links Rosenfeld's techni
cal views to Kohut's approach to the selfobject transference, as will be
discussed in chapter 6.)
Rosenfeld went so far as to indicate that verbal interpretations can
be harmful if they are made too quickly or are so persistent and rigid
that they are experienced as assaults. However, he saw a crucial role
for verbal interpretations made with tact, proper timing, and sensitiv
ity to the patient's anxiety regarding awareness of the self-other dis
tinction. If these conditions are fulfilled, the patient feels "held" by
the interpretations in an almost physical sense. (Rosenfeld's concept
of verbal and nonverbal "holding" is closer to Winnicott's notion of
the analyst as the "holding environment," to be discussed in chapter
4, than to Klein's concept of early intervention.) While verbal inter
pretations play a crucial role in Rosenfeld's treatment model for the
psychoses, their beneficial impact is a function of their ability to effect
the positive therapeutic bond not experienced in the child-mother
relationship. The resolution of the pathology still comes largely from
the verbalization of unconscious material, but Rosenfeld's view of the
treatment process includes recognition of an intense interpersonal
process that is experienced but not interpreted verbally for some
time. This conceptualization of the treatment process is based on
the developmental view of psychosis as originating in the mother's
failure to allow the child's affects inside her. According to Rosenfeld
(1987), the psychotic patient needs to "find space inside the mother/
analyst's body" more than he needs the breast (p. 278). The patient's
needs for safety and acceptance must be provided for by the analyst's
behavior in order for the child/patient to feel life. Only when the psy
chotic patient feels safe "inside" the analyst is it possible for verbal
interpretations to have meaning.
One can see that Rosenfeld departed from Klein's principles of
technique. He gave a prominent place in the treatment of psychosis to
noninterpretive intervention, principally holding and verbalizing the
patient's projectively identified affects. In fact, Rosenfeld believed that
because the patient communicates nonverbally, the analyst satisfies
many of the patient's needs by nonverbal behavior. Rosenfeld saw
the mutative effect of the psychoanalytic treatment of psychosis as
its ability to provide a satisfactory substitute for the unsuccessful
mother-child relationship in the child's background whereas for Klein
Melanie Klein 129
the curative factor is the same in the analysis of psychosis and neuro
sis, namely, making the unconscious conscious. This aspect of
Rosenfeld's approach has a closer affinity to Winnicott's concept of
holding and Kohut's use of empathy, as will become evident in
chapters 4 and 6, than to Klein's principle of immediate interpretation.
Segal's (1981) view concurs with Rosenfeld's emphasis on the
importance of containment of the patient's projective identifications,
but she is closer to Klein in her view that interpretations of uncon
scious material and the transference should be made early in treat
ment at the point of greatest unconscious anxiety in order to establish
contact with the patient's unconscious fantasy life. A fundamental
principle of Segal's technique is the desirability of making a transfer
ence interpretation in the first session with all patients; Segal felt that
such early intervention was of particular importance for psychotic
patients, as they have an immediate need for anxiety relief. It needs to
be underscored that Segal's early interpretations were based more on
the patient's use of projective identification as a defense in the trans
ference than on underlying aggressiveness, as was Klein's tendency.
For example, in the first session with a schizophrenic patient, Segal
interpreted that the girl had put all her "sickness" into the analyst as
soon as she entered the room and then had experienced Segal as a sick
and frightening person who would put the "sickness" back into her
(Segal, 1981). Klein would likely have made a more developmental
interpretation regarding early aggressiveness and anxiety. None
theless, Segal did interpret the primitive defense almost immediately,
whereas Rosenfeld would have accepted the patient's projection into
him, absorbing and thinking about it, rather than interpreting it in this
first session. Like Klein, Segal (1981) saw the mutative effects of the
analytic treatment of psychoses in the power of insight to integrate
split-off parts of the ego, with persistent interpretation of projective
identification, projection, and splitting due to excessive envy and
aggressiveness leading ultimately to a reintegration of the ego split by
these defenses.
Segal's (1981) approach to the treatment of the schizophrenic
patient is illustrated by her analysis of Edward, who had typical schiz
ophrenic symptoms: delusions of evil people taking over the world
and auditory hallucinations. Her approach was to interpret the
patient's feelings of being misunderstood and isolated and his fear of
involving his analyst in his madness. Edward kept Segal as the one
good, beneficent, unchanging figure in his life by withholding both
love and hate. Although he treated her as if she were a matter of indif
ference to him, Edward was unceasingly demanding that she gratify
130 Chapter 3
him, a behavior she responded to by interpreting what he wanted and

why he wanted it, but without gratifying the wish. According to
Segal's report, by the sixth month of analysis, the treatment had the
same focus as an analysis of neurosis: the understanding and working
through of the patient's characteristic fantasies and defenses. The dif
ference was that Edward's primary defenses were splitting—into the
persecutory and idealized objects—and magical denial.
A good example of Segal's interpretive technique is found in a ses
sion when Edward reported a voice saying "dreams, dreams." When
Segal interpreted that she had become the internal persecutor nagging
him for dreams, Edward produced a dream from the preceding night
that led to a confirmation of Segal's interpretation. In a part of the
dream a white man was turning brown because of how he had been
photographed. Segal interpreted the dream to mean that the analyst
was making Edward evil, magically turning him into feces, as he
believed he had done to his parents and to her, by wishing to watch
her in sexual intercourse. According to Segal's interpretation of the
material, Edward believed that she was retaliating by persecuting him
with demands for dreams while he reprojected by filling her with
feces, an act to which she responded by looking at him and thereby
filling him with feces. In Edward's transferential projective and intro-
jective cycles, he and his analyst were symbolically putting excrement
into each other by looking at each other. In the following session,
Edward felt much better. Segal (1981) concludes, "Obviously, he felt
that he had projected the illness into me, so that he was free, but I
became both the anxious and ill person and the external persecutor"
(p. 114).
Segal indicated that after one year of analysis the patient's delu
sions had disappeared and that he was in contact with reality and was
leading an apparently normal life; in fact, Edward was still preoccu
pied with fantasies of damage done to the earth, but he was respond
ing to such fantasies with successful concrete action on environmental
problems. Schizophrenic mechanisms continued to operate for a time,
but they were interpreted until, as Segal reports, Edward became "free
to accept a real good object, a real good experience and a real good
hope of growing out of babyhood" (p. 117).
It can be seen from this case that despite the patient's obvious dis
tress and deteriorated condition, Segal eschewed reassurance and
support. Nonetheless, she departed from Klein, who would have
offered immediate genetic interpretations of symptoms, by confining
her early interventions to the patient's isolation, desire to be under
stood, conflicts around projective identification, and wish to and fear
Melanie Klein 131
of driving the analyst mad. Segal's approach provides more of a sense

of process from beginning to middle to end than one tends to find in
the case material reported by Klein herself. We have seen that
Rosenfeld was careful to wait and utilize noninterpretive containment
before offering verbal interpretations of unconscious material. Since
Segal is perhaps Klein's closest follower, it is safe to conclude that
Klein's followers have tended to adapt her technical concepts to an
analytic approach that times interpretations to the phase of the ana
lytic process and moves gradually to deeper unconscious material.
This case also illustrates the continuity of interpretive focus in the
Kleinian tradition once the unconscious material was reached.
Similar features can be seen in the Kleinian approach to the treat
ment of borderline disorders. As we have seen, the Kleinians tend to
view the pathogenesis of this disorder in the failure of the mother to
contain the infant's projective identifications leading to the child's
overdependence on projective identification, with the consequent loss
of self to the other. The anxiety-driven reintrojection of the unwanted
parts of the self forms the primitive superego (Rosenfeld, 1978). In the
treatment situation this process repeats itself and becomes the trans
ference psychosis, the resolution of which is the fundamental issue
in the Kleinian treatment of borderline disorders. Patients believe
the analyst is hypercritical and attacking, and they counterattack.
According to Rosenfeld (1978), the borderline patient's intense projec-
tive-introjective cycles lead to confusion between patient and analyst.
Borderline patients reverse roles, attacking the analyst for possessing
the unwanted parts of the self. For example, in erotic transferences
patients believe the analyst loves them. It is the analyst's task to
accept, absorb, and put into words all such projections, as the mother
failed to do in infancy. The analyst must continually search himself or
herself to sort out all the patient's projective identifications in order to
differentiate the patient's feelings from his or her own, which the
patient cannot do. This task can be burdensome, making the treatment
of the borderline emotionally draining for the analyst. Indeed, Rosen
feld viewed many of the frustrations and treatment impasses with
such patients as results of the analyst's failure to see that his or her
frustration with the patient is a communication from the patient that
the latter has no other way of getting across. Only when analysts
are able to use their own affective reactions in the formulation of
interpretations can they effectively resolve the transference psychosis.
Rosenfeld (1978) provided a telling illustration of his technical rec
ommendations with the borderline patient in an account of his treat
ment of a multitraumatized man who was in analysis for two and
132 Chapter 3
one-half years when he displayed a hostile posture and criticized

Rosenfeld for being unceasingly critical of him. The attacks became so
vituperative that Rosenfeld felt like a helpless child filled with futile
rage. When the patient decided to leave treatment, Rosenfeld had him
sit up and go over all his criticisms. In response, Rosenfeld "did not
give any interpretations and adopted an entirely receptive, empathic,
listening attitude to him. I also examined, as much as possible, my
countertransference" (p. 219). The patient decided to stay in treat
ment, and it became clear to Rosenfeld that the patient had projected
his perception of his mother, who could not "hold" him, onto his ana
lyst and had turned the analyst into own severe, hypercritical super
ego. This interpretation began the resolution of the transference
psychosis. When the patient went through a second bout of transfer
ence psychosis some time later, it did not last long because Rosenfeld
was aware of the patient's need to project his primitive superego into
him and to communicate by reversing roles and having the analyst
feel as he had felt as a child. In this clinical illustration, too, one can
see Klein's concepts of projective identification and the projective and
introjective cycles as the core of the interpretive content, yet the
emphasis on containment, empathy, and sensitivity is a critical com
ponent of the treatment representing a clear departure from Klein's
principles of technique.
Klein's followers also have specific treatment recommendations for
the analysis of narcissistic pathology. Rosenfeld's and Segal's treat
ment approaches follow directly from their formulations of pathologi
cal narcissism as a defense against envy and excessive aggressiveness.
The crucial component of the treatment of pathological narcissism for
both theorists is the interpretation of narcissistic withdrawal and the
posture of superiority as massive character defenses against envy and
the desire to destroy.
However, the treatment approaches of Rosenfeld and Segal diverge
at this point. Rosenfeld (1971) tended to emphasize self-object fusion
as a defense against the awareness of envy and destructiveness. He
believed that in the treatment process the patient is inevitably forced
to become aware that the analyst is a separate person; this realization
results in the eruption of envy of the analyst and the wish to destroy
him or her as the patient feels humiliated by the analyst's ability to
help. The narcissistic patient becomes intent at this point on destroy
ing the analysis, now equated with the destruction of the libidinal self
of childhood dependency, the sane part of the personality. The analyst
must make a persistent effort to interpret the dependent self in order
to bring it in to contact with the analyst and, eventually, other positive
Melanie Klein 133
objects; meanwhile, the destructive omnipotent self is interpreted in

an effort to deflate it by exposing its infantile nature. When the libidi
nal self is contacted, the patient begins to form new, positive object
contacts, which strengthen it and counteract the efforts of destructive
omnipotence to dominate the personality. One can see that although
Rosenfeld viewed pathological narcissism as a defense against envy
and destructive desires, he felt that the mutative effects of the analytic
process lie more in making contact with the split-off libidinal child
hood self than in analyzing envy.
By way of contrast, Segal (1983) placed her treatment focus more
exclusively on the analysis of envy. Following Klein closely, Segal con
ceptualized pathological narcissism as a withdrawal that is based on
excessive use of projective identification and leads to fear of object con
tact. Her view was that as envy and destructive desires are analyzed,
the need for narcissistic defense and the use of projective identification
dissipate, and the patient can emerge from narcissistic withdrawal. As
such patients make object contact, they can begin to grapple with
depressive position conflicts, since both the positive and negative
aspects of the transference can now be experienced. Segal disputed
Rosenfeld's notion of a dependent libidinal self that must be contacted.
From the treatment principles advocated by Klein's followers in the
analysis of psychotic, borderline, and narcissistic pathology, one can
see a strong emphasis on projective identification, which they have
modified to an interpersonal concept in forming analytic technique.
Racker (1968) also made major contributions to this conceptual shift in
modern Kleinian treatment by pointing out that the transference is
revealed in the patient's attitude toward the analyst's interpretations.
Racker contended that patients' attitudes toward the breast extend
beyond envy and manifest themselves in a wide variety of reactions to
interpretations, including indifference, frustration, hatred, guilt,
greed, and fear of being controlled, among others. According to
Racker, whatever manner the patient adopts in relation to the ana
lyst's interpretations is a transference reaction reflecting the early atti
tude toward the breast. In turn, the analyst will have some type of
reaction to the patient's response inasmuch as every element of trans
ference is responded to with some type of countertransference reac
tion. The analyst's responsibility is to be aware of this law of the
psychoanalytic process. Only awareness of countertransference
responses can ensure that the analyst will not reenact the childhood
interaction unwittingly. Patients' deepest conflicts can be resolved by
analyzing their relationship with the analyst's interpretations if the
analyst can become aware of his or her countertransference response.
134 Chapter 3
Racker distinguished two types of countertransference based on

two types of identification with the patient. The concordant identifica
tion is the analyst's direct identification with components of the
patient's personality; it becomes the concordant countertransference.
The complementary identification is the analyst's identification with
an object of the patient's psyche, such as the patient's superego; this
becomes the complementary countertransference. Both reactions are
common in the analysis of all patients as the analyst becomes the criti
cal superego of the patient. The complementary and concordant iden
tifications indicate that the patient's transference is not simple
projection, since the analyst identifies with an aspect of the patient.
Grotstein (1986) carried this concept a step further with his view that
all projection elicits a response in the object; that is, that there is no
pure projection distinguishable from projective identification.
Even more crucial for the treatment process is Racker's distinction
between countertransference thought and countertransference posi
tion. The former is any particular discrete response evoked by the
patient, the latter a general attitude of the analyst's ego that interferes
with the analytic posture. In Racker's view, the countertransference
position is pathological because it threatens to lead to an acting-out
response rather than an interpretive one. It is the analyst's responsibil
ity to be continually vigilant of his or her responses to the patient to
prevent such an acting out. Subtle behaviors by the analyst—such as
the mind wandering—can be the acting out of a deeply unconscious
countertransference position as much as more overt acting out in
response to a frustrating and difficult patient can be.
A good illustration of Racker's use of the concept of countertrans
ference is provided by his hypothetical description of the analysis of a
patient suffering from a "neurosis of failure": the analyst's interpreta
tions of the patient's need to fail and sabotage treatment have no
impact, and the analyst feels angry and anxious over the possibility of
failed treatment. The patient fears the analyst's aggressive response,
which would be a reenactment of his childhood situation. According
to Racker, this is the most crucial aspect of the treatment process. If
the analyst acts under the threat of failure, he is dominated by his own
archaic superego object and he colludes with the patient's pathology.
However, if the analyst uses his frustration, fear of failure, and anger
to understand the internal workings of the patient, he will interpret
the transference-countertransference situation and take a step toward
helping the patient out of his cycle of failure. Psychoanalysis, for
Racker, is a continual working over of these apparent impasses and
conflicts by using countertransference responses to understand and
Melanie Klein 135
resolve pathological processes. Grotstein (1986) took a similar view of

the process by pointing out that of all the responses an analyst can
make in a given situation, countertransference will indicate which
material is most "alive" for intervention.
Ogden (1982) employed a very similar analysis of the use of projec
tive identification in the psychotherapeutic setting. In Ogden's formu
lation projective identification has three components: (1) the patient's
projection into the analyst; (2) the therapist's awareness of the pro
jected part of the self and the "metabolization" of it so that it loses the
threat it possessed when in the patient; and (3) the therapist's "giving
back" the unwanted part of the self to the patient in a way that makes
the feelings tolerable to the patient. According to Ogden, this process
is a critical component of all psychotherapy, especially with the
regressed patient. One can easily discern in this treatment theory the
developmental model of early mother-child interaction delineated by
Segal and Bion.
It should be noted that Grotstein (1986) expanded the concepts of
splitting and projective identification to all types of pathology and
normality. Grotstein pointed out that these concepts help in the
understanding not only of more severe pathology, but of neurotic
cases as well. He adopted Segal's notion that when repression is at
issue, an earlier split in the ego must have occurred. For Grotstein, the
traditional psychoanalytic defenses, such as repression, intellectual
ization, and displacement, are not necessarily indicative of pathology;
their dangerous potential arises from their use as instruments of split
ting and projective identification. It is Grotstein's contention that
unwanted parts of the self are expelled by these defensive maneuvers
and that the goal of the treatment process is to find and explicate them
in order to reintegrate the ego.
By way of summary, it can be seen that Klein's followers have con
tinued her emphasis on the pathogenicity of aggressive object rela
tions, envy, and the defenses against them, but by giving the concept
of projective identification an interpersonal meaning, they have made
significant changes in the Kleinian model of technique. The use of
countertransference to provide an analytic relationship and as a vehi
cle for interpretation has played an increasingly prominent role in
Kleinian analysis. This development has become the link between the
Kleinian school and the interpersonal theory of psychoanalysis (which
will be discussed in chapter 7). Bion (1959d) took this trend a step fur
ther with his application of projective identification to a model of group
process. In individual psychoanalytic treatment one can see that the use
of countertransference awareness has become a major instrument
136 Chapter 3
among Klein's followers in the achievement of ego integration, which

remains the goal of Kleinian treatment.
The evolution of Kleinian technical principles has led to an
increased focus on projective identification, in addition to splitting, as
the prim ary modes of defense by which the ego rids itself of
unwanted parts and weakens its cohesiveness. The goal of analytic
treatment is to reintegrate the ego by uniting the parts split off
by these defense mechanisms. Whereas for Klein ego integration
was achieved exclusively by interpretation of the unconscious, her fol
lowers have put greater emphasis on the transference-counter-
transference interaction, including the analyst's management of
countertransference responses. There are degrees of emphasis on the
analyst's containment of countertransference feelings among Klein's
followers, but even Segal, her most faithful devotee, gave a prominent
place to the "holding" of protectively identified affects in the treat
ment of severely disturbed patients before interpretation could take
place. In so doing, Segal, like many of Klein's other followers,
endorsed a model of analytic process that includes the timing of inter
pretations on the basis of the evolution of the analytic relationship, a
change in treatment that is tantamount to a revision of Klein's model
of strict reliance on depth interpretation.
These modifications in technique, along with the recognition of the
importance of early object relationships and recent spate of interest in
character pathology, have led many psychoanalytic theoreticians and
clinicians to adopt a more discerning and receptive attitude to some of
Klein's views than was prevalent in the period when Glover attacked
her work. The British "middle school" of object relations theorists,
represented in this book by Fairbairn, Guntrip, and Winnicott, fits this
category of psychoanalytic theorists who disagree with Klein's adulto-
morphic speculations on early infancy, pathologizing of development,
and interpretive style but find her object relations concepts of great
value in understanding and treating character pathology. We have
already seen the Kleinian influence on the work of Fairbairn and
Guntrip. Winnicott (1962b) was much more strongly influenced by
Klein. He drew up an impressive list of what he felt were Klein's con
tributions to psychoanalytic theory while adapting Klein's object rela
tions model to his own theory of early development and its impact on
psychopathology. It is to his views that we now turn.
CHAPTER 4
The Work of D. W. Winnicott
A lthough W in n ic o t t ' s e m p h a s is o n t h e i m p o r t a n c e o f t h e e a r l y
environment is widely acknowledged, his more specific ideas tend to

be regarded as loosely connected, defying systematic organization:
as disparate insights with little direct connection to each other. Green
berg and Mitchell (1983) present his ideas as relatively disconnected
concepts, linked only by the fact that they all deal with object rela
tions. Hughes (1987), also presents Winnicott's ideas as only vaguely
related, and Grolnick (1990), a close devotee of Winnicott, sees his
work as a conscious defiance of systematization and an embodiment
of his notion of play.
While there is certainly a strong element of play and intuition in
Winnicott's writing style, and perhaps in his therapeutic technique as
well, the view taken here is that there is a system of thought in Winni
cott's work that organizes all his ideas on development and psy
chopathology into a coherent whole. Although Winnicott did not
present his thought in a clearly organized manner, being vague and
even inconsistent at times, he espoused a consistent concept of devel
opment throughout his work, and his views on psychopathology and
treatment emanate from this developmental scheme. If one can grasp
the structure of his developmental theory, Winnicott's various clinical
papers, which may appear disparate, can be seen as different aspects
of a comprehensive view of development, psychopathology, and psy
choanalytic therapy and the individual insights which appear intu
itive and even whimsical, can be better understood as aspects of a
comprehensive theory.
To further this view of his work, Winnicott's developmental theory
will be presented first and his views on psychopathology will be taken
up in the second part of the chapter as an outgrowth of the develop
mental theory. Of all the theories discussed in this book, Winnicott's
views are most closely linked to a particular theory of development,
and it is necessary to understand this theory in order to grasp his
ideas on psychopathology and treatment. Winnicott often stated that
137
138 Chapter 4
his contribution to the psychoanalytic view of development lay

entirely in his understanding of the preoedipal phases, although he
often used his understanding of these early phases to suggest a deeper
understanding of apparently neurotic disorders. Nonetheless, Winni
cott's consistent view was that Freud's theory of the Oedipus complex
explained neurotic disorders, which he considered interpersonal; but
that the psychoanalytic theory of development needed to expand to the
preoedipal phases for an understanding of psychological organization
phases, which are at issue in more severe psychopathology and in
regressions in neurotic disorders and that must be considered in the
treatment of disorders that may appear neurotic but are refractory to
traditional psychoanalytic interventions. Consequently, Winnicott set
as his goal the understanding of the preoedipal phases of development
and their role in psychopathology.
DEVELOPMENT
W innicott's (1965) most general concepts are the "maturational

process" and the "facilitating environment." According to Winnicott,
every human organism is born with a drive, called the "maturational
process," to develop in a given direction. This constitutional given
cannot be changed, but it can be blocked if there is a failure of the
"facilitating environment" that is required for the maturational
process to take place. Maturational process and facilitating environ
ment are two sides of one coin for Winnicott, as they are for Hartmann
(1939). The environment need not be perfect, but it must be "good
enough" for the maturational process to unfold. If it is not, develop
ment is blocked and emotional disorder is the likely outcome. For
Winnicott, all symptoms are manifestations of arrests in development,
or blocks in the maturational process.
More than any other psychoanalytic theorist, Winnicott empha
sized the importance of the environment for the growth and develop
ment of the baby into a child. Since the baby cannot be thought of—
much less perform its essential task of growing into a child and
adult—without a maternal environment, the relationship of depen
dence between child and mother was the critical developmental axis
in Winnicott's thought; that is, he conceived of development as phases
of dependence of the child on the mother. Winnicott (1963a) described
three phases of dependence: absolute, relative, and "toward indepen
dence." The "transition" phase between absolute and relative depen
dence, an important developmental milestone in his theory, is a
subphase of the stage of relative dependence.
D. W. Winnicott 139
Although Winnicott wrote only one paper delineating clearly the

three specific dependency phases, he continually referred to them and
his work on development and psychopathology relies completely on
them. On many occasions he referred to shifts in dependence as the
critical features of development. Accordingly, the phases of depen
dence provide the framework for understanding his theory of devel
opment and psychopathology. By setting his various contributions
into this context, their meaning becomes more comprehensible.
It should be noted that the last phase in Winnicott's schema,
"toward independence," is given no more than a mention. He felt that
this phase, equivalent to the oedipal stage, was well conceptualized
and understood within the framework of classical psychoanalytic the
ory, and he did not attempt to contribute to this body of knowledge.
The corpus of his work focused on the preoedipal stages of develop
ment, although the distinction between oedipal and preoedipal
pathology is not so clear-cut: Winnicott believed that supposedly neu
rotic disorders often defend against more primitive issues and that
even in neurotic conditions the possibility of regressive movement is
very real. Nonetheless, Winnicott's contribution to the psychoanalytic
theory of development, psychopathology, and treatment lies in the
preoedipal phases of dependence. Given Winnicott's emphasis on
early development and the interdependence of mother and child, each
stage will be discussed from the viewpoint of both parties.
Absolute Dependence
Infant
In the first phase of development the infant is not aware of its

dependence on the environment (Winnicott, 1963a). Because the infant
cannot differentiate itself from the environment, there is no "me" or
"not me." The mother, or maternal environment, provides for the
infant's needs, but the infant has no awareness of the mother.
According to Winnicott (1960a), the infant in this phase lives entirely
in a magical world in which needs are met by their very existence.
Reality has not yet entered the infant's experience. According to Win
nicott, the infant's existence is so dependent on the mother that one
cannot speak of a baby but only of "the environmental-individual
setup." Winnicott (1952) described the situation this way: "There is no
such thing as a baby. . . . if you show me a baby you certainly show
me also someone caring for the baby, or at least a pram with some
one's eyes and ears glued to it. One sees a 'nursing couple' " (p. 99).
140 Chapter 4
There being no awareness of separateness, the infant-mother rela

tionship exists at this phase on the basis of physical contact (Winni
cott, 1963a). The infant is aware of the relationship only insofar as it
feels touched, held, or caressed or experiences other physical contact.
When such contact is withdrawn the loss of contact is felt, but separa
tion from another person is not experienced. Winnicott (1952) empha
sized here the lack of a "time factor." That is, the infant has no sense of
continuity in self or other: it does not know that it exists or that the
mother is real. Consequently, when there is separation, any physical
contact will substitute for the lost contact. This primitive sense of exis
tence is crucial for Winnicott's understanding of a variety of clinical
conditions.
Because the sense of existence is so rudimentary, any disruption
threatens the minimal sense of existence the infant is able to feel in
this stage. As there is no self that could manage anxiety as a warning
signal, all anxiety is experienced as annihilation anxiety (Winnicott,
1952, 1960a). Consequently, the infant in this phase lives on the brink
of "unthinkable anxiety" (Winnicott, 1952). Since the lack of temporal
sense gives the infant little belief in relief, annihilation anxiety can be
quickly produced. As discussed in more detail later, environmental
adaptation must be "near total" (Winnicott, 1956a) so that the small
doses of reality are manageable and annihilation anxiety is fended off.
Signals do not exist in the phase of absolute dependence, according
to Winnicott (1963a). Since the infant is unaware of separateness, it
has no awareness of a person on whom it depends and cannot signal
its needs. Thus, the caretaker must interpret the infant's behavior as
communication even though the infant has no intention of communi
cating. The infant's needs must be met without any communication
from the infant, and if the infant survives, some needs have appar
ently been met. Insofar as mentation is possible, needs appear to the
infant to be satisfied by their very existence; they are experienced
as bringing their own gratification. Thus, according to Winnicott
(1945, 1960a), the infant in this phase lives in a magical world of
omnipotence in which mentation produces gratification.
The first critical developmental task is the achievement of integra
tion. Experience is "un-integrated" in this stage, as feelings, needs,
and tension states are not experienced as belonging to a common
whole (Winnicott, 1945,1962). There is no "me" to hold together expe
riences, no sense that discrete experiences are linked and no time factor
to connect discrete experiences. Here Winnicott is drawing on Glover's
1937 concept of "ego nuclei." Experiences are not linked together in
time and are therefore not experienced as a "lived temporal unity."
D. W. Winnicott 141
According to Winnicott, they have no "unit status." Because experi

ence is discrete rather than continuous, there is no "lived psychic
reality" (Winnicott, 1945). The sense of temporal continuity is an
achievement as experiences are gradually linked together into an
integrated self.
It follows from the inability to differentiate "me" from "not me"
that the infant is not yet able to experience itself as a person. This
inability means there is a lack of "personalization" of experience; the
infant's sensations, needs, and feelings are not personalized into "my
experience." In Winnicott's words, experience is not yet "localized."
"Personalization" is the second major developmental achievement, an
outcome of successfully traversing the early phase of dependence.
From the infant's point of view, needs do not have to make a
detour through reality to be met; thus there is no reality sense. The
third primary task of absolute dependence is the gradual develop
ment of the sense of reality. As noted earlier, with each frustration, the
infant experiences a bit of reality, and its "omnipotence" is pierced.
Because "realization" is made possible by environmental failures
(Winnicott, 1945), unsuccessful adaptation is as crucial to the infant's
sense of reality as successful adaptation. Winnicott (1963d) points out
that, "there is no question of perfection here. Perfection belongs to
machines; what the infant needs is just what he usually gets, the care
and attention of someone who is going on being herself" (pp. 87-88).
The environmental adaptations and failures provide the infant with
"doses" of reality in manageable portions: "The whole procedure of
infant care has as its main characteristic a steady presentation of the
world to the infant" (Winnicott, 1963a, p. 87). As the infant's needs are
met, the nascent sense of self grows and frustrations become gradu
ally more tolerable, resulting in the ability to feel a temporal sense.
Gradually, reality and fantasy become distinguishable; however, there
is still very little sense of reality in the phase of absolute dependence.
Only with its passage into later dependence phases is the infant able
to differentiate pure mentation, as in fantasy, from the experience of
reality and thereby achieve "realization." (This conceptualization
anticipated Kohut's, 1971 view, discussed in chapter 6, that maternal
empathy and its failures are instrumental in the journey from archaic
grandiosity and idealization to a realistic view of self and others.)
As there is no self-other distinction in the phase of absolute depen
dence, there are no objects yet for the infant (Winnicott, 1971). Insofar
as one can speak of objects in the infant's awareness, they are experi
enced within the orbit of infantile omnipotence; that is, the infant feels
they are under its complete control. One cannot yet speak of object
142 Chapter 4
usage, although there is a sense of "object relating" within the

omnipotent sphere. In order to use an object, it must be seen as exter
nal to the self, and this awareness does not occur until the phase of
relative dependence.
If there is no concept of an external object, what does the infant see
when it looks at the mother? According to Winnicott (1971), the infant
looks at the mother and sees itself. The mother is looking at the baby
and "what she looks like is related to what she sees there" (p. 112).
The mother performs a "mirror role" (but this function is not the same
as Kohut's well-known concept of "mirroring," as shall be seen in
chapter 6); that is, the mother functions as a mirror by giving back to
the infant what it gives in its own look. The infant's first image of
itself includes a connection with the mother; by the mother's look, the
infant knows that it is seen.
Winnicott (1947, 1962, 1963b) adopted Freud's and Klein's view
that aggression is inborn. Winnicott's views on the infant's aggression
are somewhat confusing because he referred to both an innate aggres
sive drive and an "inherent aggressiveness in the love-impulse." Win
nicott did make clear that in the early phase aggression is without
intention, "pre-ruth." He pointed out that even in the womb, the baby
kicks; aggression is originally "almost synonymous with activity"
(Winnicott, 1950). The infant has no awareness that its aggressiveness
can hurt the other. Aggression is inherent in the "love-impulse," so
that when an infant is excited it "destroys." The excitement of instinc
tual gratification means an attack on the mother's body. Nonetheless,
the infant does not know that the attack is visited on the same object
whom it loves when not excited. As Winnicott (1945) wrote, "The nor
mal child enjoys a ruthless relation to his mother, mostly showing in
play, and he needs his mother because only she can be expected to tol
erate his ruthless relation to her even in play, because this really hurts
her and wears her out" (p. 154).
Winnicott's views on early aggression bear a similarity to those of
Klein, by whom he was influenced, and yet they are decidedly differ
ent. As we saw in chapter 3, for Klein the aggressive and libidinal
drives are directed toward different objects, since "good" and "bad"
objects are split. For Winnicott, the drives are initially fused, since
aggression is part of love, and the split is between excited states, which
always include aggression, and quiescence. Splitting of good and bad
objects is a defensive pathological reaction to the failure of the environ
ment to hold the early aggression. If the environment is good enough,
the infant has no awareness that its aggressiveness can injure and the
aggression therefore remains a part of the "love impulse."
D. W. Winnicott 143
There comes a point at about six months of age when the infant
becomes aware that it does not control gratification. This critical
developmental step marks the beginning of the self-object distinction.
The infant is now starting to see objects as outside the self, marking
the entrance of the reality principle into the infant's life (Winnicott,
1971). According to classical psychoanalytic theory, the infant, out of
frustration, has an aggressive response to the awareness of reality.
Winnicott (1971) saw a more constitutive role for aggression—in the
development of the infant's very sense of reality. He believed that the
infant experiences its aggressiveness as expelling the object from the
sphere of omnipotence. The object must be "destroyed," and only
later, when it is "refound" as an external object in the phase of relative
dependence, can it be "used." Thus, for Winnicott, aggression plays a
crucial role in the development of the self, the sense of reality, and the
recognition and use of objects.
To summarize the phase of absolute dependence from the view
point of the infant: the infant lives in a world of magical omnipotence
in which there is no sense of self or reality, and no experience of
objects. Experience is discrete, disconnected, fleeting, and outside of a
"lived psychic reality." All these aspects of the absolute dependence
phase are emphasized in Winnicott's work because he felt that they
provide clues to the mystery of primitive psychopathology.
Mother
Winnicott (1956a) believed that the mother goes through a phase of

"primary maternal preoccupation" beginning in the later stage of
pregnancy, when she becomes emotionally focused on the soon-to-be-
born child. This emotional preoccupation includes a natural adapta
tion to give up her previous life entirely or partially in order to meet
the needs of the infant. All the preparations made for the coming child
reflect a new state of mind, that Winnicott considered so extreme he
would regard it as an illness if there were no baby.
The mother's absorption with the infant shortly before and for
some time after birth allows the infant to bond without awareness of
separateness and thereby makes possible the infant's experience of
omnipotence. For the mother, normal activities are limited, and her
emotional life is focused on the new baby.
Because the infant in the phase of absolute dependence cannot sig
nal its needs, the mother must rely on empathy to meet them.
Through her identification with the infant, based on her own experi
ence of being mothered, the mother is able to know what the infant
144 Chapter 4
needs without being told or signaled. For Winnicott, empathy does

not mean understanding verbal communications; it does not even
mean grasping their affective undertones. It means knowing what the
infant needs, even though the infant cannot communicate its needs.
The empathic mother interprets her infant's behavior as a communica
tion to her, even though the infant does not intend to communicate.
The infant cries and the mother interprets the cry as a signal of hunger.
The infant has not signaled, but its survival depends on the mother's
ability to interpret its behavior as a signal. This definition of empathy
is to be contrasted with Kohut's contemporary usage, described in
chapter 6. Later, in the phase of relative dependence, communication
between infant and mother enters their relationship, but in the phase
of absolute dependence empathy through identification is the primary
mode of adaptation the mother must make.
The mother need not adapt perfectly to the infant's needs, but the
adaptation must be "good enough" to allow the infant the experience
of omnipotence (Winnicott, 1960a); the infant is not yet ready for real
ity, and too early a confrontation with it would be dangerous. Adapta
tion by the mother must be "near absolute" so that the infant can live
in its delusional world of omnipotence, however, it cannot and should
not be perfect because environmental "failures" are necessary to
prepare the way for the infant's descent into eventual reality.
The mother's provisions for the infant fall into two categories of
maternal functions: the "object mother" and the "environmental
mother" (Winnicott, 1963b). The object mother provides for the "object
instinctual needs" by satisfying hunger, holding the infant, and keep
ing diapers clean. However, the object mother alone is insufficient.
The infant has "ego needs" from the very beginning, and these are not
met by the object mother. For example, the infant needs an environ
ment in which air and water temperatures are relatively comfortable,
the noise level is neither unstimulating nor assaultive, and there is
visual stimulation that is interesting without being overwhelming.
Most im portantly, the environm ent must be relatively free of
"impingements" that would interfere with emotional growth.
Winnicott placed great emphasis on ego needs because he felt that
classical psychoanalytic developmental theory overemphasizes
"object instinctual" needs while failing to recognize the importance of
the early environment and overall maternal care in the growth of the
ego. Because the phase of absolute dependence is subsumed under the
oral phase in classical theory, emphasis is placed on the meeting of
"object instinctual" needs. Winnicott pointed out that the primary
developmental tasks of this phase—the development of "realization,"
D. W. Winnicott
personalization, and integration—are all ego needs and that it is possi

ble for "object instinctual" needs, such as hunger, to be met ade
quately while developmental needs are not. Under such conditions
the meeting of "object instinctual" needs does not lead to growth; it is
only when the ego needs are met that object instinctual fulfillment
promotes personality growth. For Winnicott, then, ego needs take
precedence over object instinctual needs.
Winnicott's (1960a) general term for the role of the environmental
mother in this phase is "holding," a term he described as follows:
The term "holding" is used here to denote not only the actual physical
holding of the infant, but also the total environmental provision prior to
the concept of living with. In other words, it refers to a three-dimen
sional or space relationship with time gradually added.... It includes the
management of experiences that are inherent in existence, such as the
completion (and therefore the non-completion) of processes, processes
which from the outside may seem to be purely physiological but which
belong to infant psychology and take place in a psychological field,
determined by the awareness and empathy of the mother [pp. 43-44].
The infant's needs and emotional states are "held" well enough by the
good enough mother, who is able to contain whatever tension states
occur by her empathy.
Because the only anxiety the infant experiences in this phase is
annihilation anxiety, the mother must "hold" the anxiety and meet the
infant's needs well enough to keep the infant from experiencing a
sense of annihilation. To succeed in this preventive task, the maternal
environment must protect the infant from impingements. According
to Winnicott (1960a):
The holding environment therefore has as its main function the reduc
tion to a minimum of impingements to which the infant must react with
resultant annihilation of personal being. Under favorable conditions the
infant establishes a continuity of existence and then begins to develop
the sophistication which makes it possible for impingements to be
gathered into the area of omnipotence [p. 47].
The principal function of "holding," then, is keeping the infant suffi

ciently free of environmental impingement that it lives within its
fantasy of omnipotence and continues to grow.
Included in the emotional states of the infant that must be "held"
by the "environmental mother" is the infant's aggression (Winnicott,
1950). Recall that, according to Winnicott, the infant's erotic need has
146 Chapter 4
an inherent aggressive "pre-ruth" component. The infant's object

instinctual need "destroys" by "attacking" the mother's body and the
mother must be willing to absorb and hold such aggressive attacks. As
we have seen, Winnicott (1971) believed that the destructive drive
"pushed" the object outside the sphere of the infant's omnipotence,
thus "objectifying" the object for the first time. It is crucial that the
infant experience the object's survival of its destructive attacks; only
then can the infant turn a "subjective object" into an "objective object."
That is, only when the object survives the destructive attack can it be
"used" by the infant. There is an inherent connection, for Winnicott,
between the object's ability to survive the destructive attacks, the
infant's growth out of omnipotence, and its ability to use an object.
The mother's ability to hold the destructive attacks and not retaliate is
critical to the infant's ability to perceive reality, see objects as "not
me," and use objects as separate others for further growth.
If the environment is able to perform its functions well enough in
this phase—that is, if it is able to prevent impingements; hold the
infant's frustrations, thereby preventing annihilation anxiety; "hold"
the aggression; and meet the infant's overall ego and id needs—the
infant will begin to experience a sense of continuity of its various need
states and a rudimentary sense of integration, personalization, and
realization will occur. The inevitable environmental failures will force
some degree of reality on the child, but these moments of brief recog
nition will not disturb infantile omnipotence because the holding has
been "good enough" for the child to feel that its needs bring their own
gratification. The growing child is then ready for what Winnicott sees
as the first major developmental task in the formation of the self: the
awareness of dependence.
Relative Dependence
Child
Many investigations of early development have noted that at six to

eight months of age infants tend to become aware of their surround
ings in a new way (Spitz, 1965; Mahler, 1975; Stern, 1980). They reach
out to explore with their hands, especially the faces of caretakers, and
they often do to the mother what the mother does to them (Mahler,
1975): they touch, soothe, caress, and stroke the mother's hair; some
even offer the bottle to the mother. The infant for the first time differ
entiates clearly between mother and others and between familiar fig
ures and strangers. Mahler (1975) called this newfound awareness
D. W. Winnicott 147
"hatching," as the infant is now able to differentiate itself from

the environment.
For Winnicott (1960a, 1963a), the critical aspect of this develop
mental shift is the child's awareness of its dependence on an object
outside itself. It is by virtue of this awareness that the infant is said to
move from absolute to relative dependence. The infant now becomes
interested in exploring the object of dependence but also becomes
anxious in separating from it. The first awareness of "me" versus
"not me" is shown by the infant's reaching for objects, a new kind of
play. "We can say at this stage a baby becomes able in his play to
show that he can understand he has an inside, and that things come
from outside" (Winnicott, 1945, p. 148). The infant recognizes that the
environment is separate, "out there," and yet necessary to meet its
needs. The infant becomes aware in a rudimentary way that its needs
do not automatically entail their own gratification. This awareness of
separation from the environment is the beginning of the sense of
self—and, simultaneously, an experience of loss.
Winnicott (1963a) pointed out that infants now experience a new
kind of anxiety: "When the mother is away for a moment beyond the
time-span of his (or her) capacity to believe in her survival, anxiety
appears, and this is the first sign that the infant knows" (p. 88). Spitz's
research shows that around this time all infants experience both sepa
ration and stranger anxiety, with separation anxiety occurring about
one month earlier. There are wide variations in degree, exact time of
onset, and duration for both forms of anxiety, but Spitz was able to
show, by slow-motion filming, that all infants experience both forms,
in however mild or transient a form. This finding supports Winni
cott's point that both anxiety reactions occur in response to the
infant's awareness of its separateness from the maternal environment.
The infant attempts to preserve the attachment by identification
with the mother. The infant's imitation of the mother which begins at
this time, is conceptualized by Winnicott (1963a) as the first identifica
tion and as occurring in response to the newly discovered gulf
between itself and its mother.
Relative dependence, for Winnicott, encompasses an extended
developmental process from the breakdown of the omnipotence of the
absolute dependence phase to the acceptance of reality and ambiva
lence toward whole objects. Included in this phase is a variety of
developmental tasks, the first of which is to manage the anxiety of
separation from the mother, the awareness that it cannot meet its own
needs. To master this anxiety and bridge the transition to reality orien
tation, the infant utilizes a variety of possessions and experiences
148 Chapter 4
referred to by Winnicott as transitional phenomena, the "transitional

object" being only one, albeit the best-known, example (Winnicott,
1951). Winnicott's concept of transitional phenomena is so widely dis
cussed, so commonly oversimplified and misunderstood, and yet so
important to his thought that it is worth considering in detail.
The essence of any transitional phenomenon, for Winnicott, is its
function as an intermediary between the fantasied world of omnipo
tence in the absolute dependence phase and the acceptance of reality,
which is the outcome of the phase of relative dependence. It is not
commonly recognized that Winnicott (1951) distinguished three types
of transitional phenomena. Initially, anything that is both "mine" and
"not mine" is a transitional experience. Cooing serves this function;
the infant emits the sound from within but hears it from without. The
experience helps the infant differentiate "in" from "out" while provid
ing a connection between them. Thumb sucking is a different type of
transitional phenomenon, requiring organized control over the mus
culature; using the thumb is a developmental advance in that it begins
the substitution of limited muscular control for the fantasy of omnipo
tent control. Thumb sucking indicates a willingness to yield the fan
tasy of omnipotence as the exclusive means of getting needs met and is
a further step toward the acceptance of reality, even though the reality
in this instance is a part of the infant's own body.
Some time after the thumb is used, the infant will draw from the
realm of transitional phenomena a single object outside itself to which
it becomes closely attached, its first "not me" possession. It is this pos
session that is called a transitional object. The infant knows that the
stuffed animal, blanket, toy, or doll is outside of itself; therefore, it has
relinquished to an even greater degree than with thumb sucking the
fantasy of omnipotence of the absolute dependence phase. Not only
does the use of the first possession substitute muscular control for
omnipotent control, but the infant is now using something it knows
not to be itself. The possession can be lost or misplaced, and it must be
found. By using an object outside its omnipotent control and outside
itself, the child has taken a further step toward reality acceptance:
It is true that the piece of blanket (or whatever it is) is symbolical of

some part-object, such as the breast. Nevertheless the point of it is not
its symbolic value so much as its actuality. Its not being the breast (or
the mother) is as important as the fact that it stands for the breast
(or mother) [Winnicott, 1951 p. 233].
On the other hand, the transitional object is not a reality object. It is

D. W. Winnicott 149
imbued with intense, powerful, personal meaning. Finding the object

immediately reduces intense anxiety; sometimes anxiety can be
soothed in no other way. The object is treated as though it were the
mother, although the child knows that it is not. The paradox of the
transitional object is that it is neither real nor delusional. It is illusory,
an intermediate area of experience, lying between reality and fantasy
(Winnicott, 1951,1971; Barkin and Grolnick, 1970). According to Win
nicott (1971), the transitional object begins the world of illusion and
prepares the way for play in childhood. Child's play, according to
Winnicott, is based on giving an illusory meaning to something real.
The attributed meaning is known to be an illusion but is treated dur
ing the play as though it were not. Because the meaning must always
be personal and unique to the individual, transitional phenomena and
its later variant, play, are always creative. Winnicott felt that all child
and adult creativity, as well as aesthetic experience, are transitional
phenomena and that this intermediate area of experience must con
tinue into adult life for creative and cultural living, which he identified
with mental health.
The transitional object is a response to loss; for this reason, Winni
cott (1971) believed that the transitional object is used as a "defense
against depressive anxiety." Nonetheless, he felt that the importance
of the transitional object lies in its not being the mother or the breast;
that is, it indicates that the infant, through its ability to use a piece of the
real world to fill an emotional need, is moving away from omnipotent
fantasy and toward reality.
To realize its function fully, the transitional object must have cer
tain features: the child must have complete ownership and dominance
over it; it must contain intense feelings, both positive and negative; it
must provide the experience of warmth and cuddling; and it must
provide a feeling that it has a reality of its own. The assumption of
dominance is critical so that the infant is not unduly frustrated, yet the
object must have reality so that dominance is not confused with
omnipotence. Having all these characteristics, the transitional object is
able to be used as though it were the mother; it functions as though it
were the mother, yet the infant knows and is reminded constantly by
the object that it is not the mother.
Transitional objects help the infant ease the separation and stranger
anxiety caused by awareness of dependence and separateness.
Although these experiences are painful, they also betoken a develop
mental advance: the infant has moved from annihilation anxiety to
anxiety of object loss. Episodes of stranger anxiety end when the
mother returns, implying the existence of a maternal representation
150 Chapter 4
when she is not present. Stranger anxiety also implies a specific

attachment to the mother as someone who has continuity in the
infant's mind. Beginning with this attachment to the mother, the
infant acquires a sense of the familiar. Experiences begin to connect
with each other and the infant makes developmental advances it was
incapable of in the phase of absolute dependence.
Primary among these new advances is the sense of "temporal inte
gration" (Winnicott, 1962). Since continuity is now experienced, the
previously discrete ego nuclei begin to link together and develop into
a sense of temporal continuity, eventually leading to the achievement
of the sense of integration. The infant now has the growing sense of
itself as one person with a variety of experiences, rather than an
awareness of only discrete experiences. This sense of continuity leads
to a rudimentary sense of self.
Since experiences are now felt as belonging to a unit, a sense of
"personalization" begins to develop. The growing child begins to
experience what Winnicott calls a "lived psychic reality." The new
sense of an "internal," a "me," is not solely a cognitive formulation.
There is an "internal environment," since living now includes an
"inside" and this is distinguishable from what is "outside." Whereas
experience in the phase of absolute dependence simply occurs, in rela
tive dependence experience begins to belong to "me," and it is in this
phase that the child begins to use the word "me."
Personalization is closely linked to the third achievement in the
development of the self in the phase of relative dependence: the
growing sense of "realization." The infant has been "hatched" from
the fantasy world of omnipotence into the realization that the envi
ronment is separate from the personalized self. The awareness of this
distinction is the origin of the infant's sense of reality. In the fantasy
of omnipotence, there is no sense of the reality of the world or of the
self. As the infant gains a sense of personal reality, the awareness of
the self-other distinction takes place and the world is experienced for
the first time as having a reality apart from the infant's mentation.
This awareness culminates in the sense of the self as a real person and
the sense of the world as a separate reality. Both world and self are
becoming "realized."
The infant now knows that its needs are met by a mother outside
itself. The mother, as the medium for the fulfillment of the need,
begins to represent reality to the child. In fact, Winnicott says, the
mother "brings reality to the child" via her position between the need
and its fulfillment. Every delay, every imperfect meeting of need,
forces reality upon the infant who becomes aware that in order to
D. W. Winnicott 151
have any control over the meeting of its needs, it must signal to its
provider. The infant now intends to communicate. Whereas in the pre
vious phase the infant cried and the mother interpreted the cry to
mean hunger, now the infant cries to tell the mother of its hunger or,
at a more advanced level of this phase, the child points to food. Now
the mother need only understand communication; she does not have
to create it. The child does not rely on the experience of need or wish
to bring gratification; it relinquishes the magical control of pure men
tation for the reality-based control of gesturing and communicating
through word and deed.
Now that temporal integration begins to occur, the infant can
"carry" experience from one contact to the next. As it begins to form a
sense of a single person who meets its needs, the growing child begins
to experience a relationship. And once the sense of the mother is
maintained without her physical presence, an "ego-relatedness" takes
place between mother and child. Out of this psychological way of
relating, the object is internalized and a relationship is formed.
According to Winnicott (1956d), the infant is now capable of a rela
tionship based not on physical contact but on the psychological recog
nition of the other. The "ego relationship" between infant and mother
is the key to the development of the infant's concept of the other as a
separate person. In this sense, Winnicott (1956a) opposed the tradi
tional analytic conception that recognition of others as separate from
the self is a product of frustration: "From this angle the recognition of
the mother as a person comes in a positive way, normally, and not out
of the experience of the mother as the symbol of frustration" (p. 304).
The other side of ego relating is the development of the ego orga
nization. "The first ego organization comes from the threats of
annihilation which do not lead to annihilation and from which,
repeatedly, there is recovery. Out of such experiences confidence in
recovery begins to be something which leads to an ego and to an ego
capacity for coping with frustration" (Winnicott, 1956a, p. 304). The
infant develops an ego organization as he or she internalizes his ego-
relatedness to the maternal environment that protects him outside of
his awareness.
For the development of the capacity for such ego relatedness to
take place, the infant must learn to be alone. First, the infant must be
able to be alone in the presence of the mother, a paradox Winnicott
noted, and even emphasized. By being alone in the presence of the
mother, the infant experiences a relationship without physical contact
or need gratification. The environmental mother is providing the child
with an ego relationship, thereby allowing for her internalization. In
152 Chapter 4
Winnicott's view, once the mother is internalized, the infant can be

alone without her. Ego-relatedness and internalization imply and fos
ter each other, and as they develop the temporal integration of the
personality is strengthened (Winnicott, 1962).
At this point, "holding" has been replaced by "living with." The
infant is developing a sense of self and the ability to communicate its
needs to the mother. Recognition by the communicating infant that
the mother is a separate person to whom communication must be
made indicates that a relationship, a "living with," is beginning to
form. The physical correlate of this shift from "holding" to "living
with" is the difference between the mother holding the child in her
arms and the child slipping off her lap, crawling away, and appealing
to the mother when help is needed. However, the defining feature of
the shift is not so much physical growth as the infant's recognition of
the mother as a separate person to whom it communicates and with
whom it begins to form a relationship (Winnicott, 1960a).
With the development of the temporal unity of the personality,
the growing child links experiences and begins to perceive that the
"object mother" whom it "destroys" when excited is the same as the
"environmental mother" whom it loves when quiescent. This inte
gration of objects creates a special problem, as the infant is now
aware that it can injure the object of its love (Winnicott, 1963b). The
infant has shifted from the "ruthlessness" of object instinctual
attacks on the mother's body to awareness that it has been attacking
the mother who cares for it.
Although this formulation calls to mind Klein's "depressive posi
tion," Winnicott (196b3) preferred the expression "stage of concern."
In Winnicott's formulation, the child's concern is rooted in its belief
that it can injure the mother with its aggressiveness but this concern
need not lead to depression, nor even "depressive anxiety." Like
Klein, Winnicott believed that guilt originates with the experience of
object integration (in opposition to Freud, 1923 who dated guilt from
the oedipal phase). The realization that the child has hated or even
been angry at the same person who is loved and depended on pro
duces anxiety over injuring the object, and this anxiety inevitably
leads to guilt (Winnicott, 1954a). However, for Winnicott, unlike
Klein, the development and vicissitudes of guilt may lead either to a
healthy concern for others or to depressive anxiety, depending on the
environmental reaction to the child's aggression.
The mother's reaction to her child's aggression will be discussed in
detail in the next section; here we note that the crucial experience from
the infant's viewpoint is that its aggression neither injures nor
D. W. Winnicott 153
destroys its mother. If the infant sees that its mother survives and
accepts its aggressiveness, it is able to believe in the value of its
aggression, which can then remain connected with the "erotic
impulse." Instead of linking aggressiveness with destruction, the
infant believes its aggression promotes a connection to the maternal
love object: "On the erotic side there is both satisfaction-seeking and
object-seeking, and on the aggressive side, there is a complex of anger
employing muscle erotism, and of hate, which involves the retention
of a good object-imago for comparison" (Winnicott, 1963b, p. 74). The
child is now able to experience love and hate, positive and negative
feelings, to the same mother without fear of injury or destruction.
Guilt, in the sense of concern for others, is felt, but it facilitates rather
than blocks the maturational process by leading to altruism and ethi
cal responsibility. In Winnicott's view, the experience of ambivalence
toward whole objects is the origin of the adult desire to contribute to
family and community.
The experience of love and hate toward the same object makes a
crucial contribution to self development. First, the mother is perceived
as a whole object, which furthers the "m e"-"not me" distinction.
According to Winnicott (1963b), as the infant experiences ambiva
lence, "the infant is beginning to relate himself to objects that are less
and less subjective phenomena, and more and more objectively per
ceived 'not-me' elements" (p. 75). Secondly, the child's experience of
self is enhanced: as the mother becomes a whole object, "inner psychic
reality...becomes a real thing to the infant, who now feels that per
sonal richness resides within the self. This personal richness develops
out of the simultaneous love-hate experience which implies the
achievement of ambivalence, the enrichment and refinement of which
leads to the emergence of concern" (p. 75).
The fantasy of injuring the loved object is, according to Winnicott,
the first experience of agency. In this regard, Winnicott continued
to follow Klein. In the phase of absolute dependence the infant is
helpless and experiences itself as a victim of attack; in the phase of
relative dependence the infant experiences itself as an agent of
aggressive attack. The successful completion of this phase results in
a sense of responsibility and the capacity for guilt, with minimal
anxiety over aggressiveness. The sense of agency and of its corollary,
responsibility, culminate in the initial sense of an active self able to
control life events.
If all goes well, toward the end of the phase of relative dependence
the child has an integrated sense of self as continuing over time and
differentiated from others. The child is now at the whole object level,
154 Chapter 4
able to contain ambivalence and therefore able to maintain relation

ships with others. A child whose development has proceeded to this
point has the capacity for ambivalence required for entry into the
oedipal phase.
The Mother
The entire developmental process through the phase of relative

dependence depends on the "good enough" maternal environment.
Both the "object mother" and "environmental mother" must make
shifts in their adaptations in accordance with the child's developing
sense of self. In the phase of absolute dependence the mother made a
near-total adaptation to "hold" the infant; now the mother must begin
to develop her own growing independence in order to "live with" the
child. According to Winnicott (1960a), one of the mother's primary
tasks in this phase is to "allow" the child to communicate and then to
respond to the communication. She responds less out of empathy and
more out of an understanding of the child's communication: "there is
a very subtle distinction between the mother's understanding of her
infant's need based on empathy, and her change over to an under
standing based on something in the infant or small child that indicates
need" (p. 51).
The mother who continually utilizes maternal empathy after the
infant has begun to recognize the self-other distinction is fostering a
continued sense of omnipotence in the infant at the expense of its
growing need for a sense of self and movement toward reality. The
mother must begin to develop a sense of independence from the
infant corresponding to its movement away from her so that the infant
can begin to relinquish the fantasy of omnipotence in favor of reality,
develop a sense of self, and form a relationship to a separate other.
The mother's increasingly imperfect adaptation allows the child to
experience gradually the delays and imperfections of reality. Her
"failures" are critical in "bringing reality to the child" (Winnicott,
1956a). She allows "real-ization" by meeting needs imperfectly and
by her own need to become more independent. The mother's near
perfect adaptation now becomes an environmental impingement,
just as inability to adapt was an impingement in the previous phase
(Winnicott, 1960a).
Along with her growing independence, the mother must develop a
different type of relationship with the child. She must give "ego sup
port" to the child's growing need for independence both by providing
a safe, protected environment, and by allowing the child to explore,
D. W. Winnicott 155
manipulate the environment, and to play without interference while

continuing a relationship with the child. According to Winnicott
(1958), since the child needs a relationship without dependence on
physical contact, the mother must offer such an "ego relationship" in
which emotional contact is provided without instinctual gratification.
We saw earlier that Winnicott emphasized the importance of the
child's being alone in the presence of the mother in the develop
ment of the sense of self. The mother's role in the development of
the capacity to be alone is to offer an ego relationship in which she
is not threatened by the child's being alone in her presence and she
does not withdraw. She must allow herself to be ignored. Her pres
ence, experienced by the child, provides an ego relationship that
can now be internalized.
The mother's willingness to allow the child to be alone in her
presence is one special aspect of her provision of "ego coverage"
(Winnicott, 1960a). This creation of a safe, protective environment, rel
atively free from major disturbances, allows the child to "go on being"
and fosters spontaneous ego growth (Winnicott, 1963b). If the child
must be concerned about disturbances, its emotional focus is diverted
from natural growth and development to reactions to the environ
ment. Such impingements and the child's reactions to them lead to
preoedipal psychopathology.
Toward Independence
Once the child has an integrated sense of self and has internalized
whole objects, it is prepared for movement Winnicott calls "toward
independence," synonymous with the oedipal phase of development.
He proffered no contribution to the classical psychoanalytic literature
on the oedipal period and, in fact, wrote little regarding the phase of
"toward independence," as he felt his contribution to psychoanalytic
thought was in his ideas about the earlier phases. Winnicott's rela
tively sparse comments regarding the phase "toward independence"
are summarized briefly to provide closure to his developmental theory.
As the child moves toward independence, its sense of self relies
even less on physical needs and sensations than in the previous phase.
Now that the child has a sense of "an internal environment" it not
only is able to be alone but often seeks to be alone.
Winnicott adopted the classic analytic position that the child is now
able to manage three-party relationships. This developmental step has
the same meaning for Winnicott that it has in classic psychoanalytic
theory: the child can now move beyond an independent relationship
156 Chapter 4
with each parent to an awareness that the parents have a relationship

with each other from which the child is excluded. In Winnicott's for
mulation, as in classical theory, ambivalence toward the parents and
their relationship is crucial to the resolution of the oedipal phase.
Winnicott believed that the child's developed self and ability to expe
rience ambivalence allowed movement from the child's "living with"
each parent to child, mother, and father "living together." The child
becomes part of a network of others, able to relate to others as part of
a family or other unit.
The mother must proceed with her movement toward a more inde
pendent life. The child now fits into her life rather than being the
almost exclusive focus of her life as in the previous phases. The
mother's achievement of independence while maintaining her rela
tionship with the child and meeting such dependency needs as the
child has allows the child to experience living together in the family
unit. The mother is now able to reinvest in the relationship with the
father in a more focused way. She may return to work, have other
children, begin a new career, or engage in other activities, new or old.
PSYCHOPATHOLOGY AND TREATMENT
All that has been said thus far regarding preoedipal development
assumes the presence of a "good enough mother" who provides a
"facilitating environment." What happens if the maternal environ
ment is not "good enough" and impingements occur? Winnicott's
answer is that the effects of environmental failure depend on the
developmental phase and the type of impingement. Although Winni
cott did not set forth a comprehensive system of psychopathology, he
did conceptualize impingements and their effects sufficiently so that
one can discern in his work a theory of clinical syndromes and their
causes. It should be noted that he believed certain specific impinge
ments and their sequelae in a preoedipal phase could lead either to a
clinical syndrome or to a more transient manifestation in a higher
level neurosis. Because this difference was not systematized in his
work, it has never been clear why certain impingements eventuate in
relatively fixed syndromes in some cases and transient manifestations
in others.
Winnicott's view of psychopathology follows his developmental
scheme so closely that he felt the words patient and baby could be used
interchangeably. All psychopathology, in Winnicott's view, results
from an insufficiently facilitating environment, that causes the infant
or child to react to environmental impingement, and thus arrests
D. W. Winnicott 157
the maturational process. Impingements in the phase of absolute

dependence potentially result in the most disabling forms of emo
tional disorder because they interfere with the most basic psychologi
cal structures. In the phase of relative dependence impingements may
create character pathology whereas oedipal phase impingements are
the source of neurosis.
Psychopathology and Absolute Dependence
Winnicott (1952) called impingements in the phase of absolute

dependence "privation" because needs are not met. "Deprivation,"
which signifies that something that was given is subsequently taken
away, applies to the phase of relative dependence. In the phase of
absolute dependence, if the satisfaction of environmental and object
needs is not "good enough," the child cannot focus on "going on
being." If maternal empathy is lacking, omnipotent fantasies are
assaulted by reality before the child is ready to relinquish them and
accept reality. The child becomes aware that needs are not met by
their existence and is forced into a premature awareness of self-object
differentiation. This premature awareness interrupts the development
of integration, personalization, and realization. All forms of psy
chopathology originating in the phase of absolute dependence result
from this dynamic.
Premature awareness of the self-other distinction results in the
eruption of annihilation anxiety, an "unimaginable terror" or fear of
falling apart, akin to the anxiety of falling endlessly through the air
(Winnicott, 1952). "Maternal failures produce phases of reaction to
impingement and these reactions interrupt the 'going on being' of the
infant. An excess of this reacting produces not frustration but a threat
of annihilation. This in my view is a very real primitive anxiety, long
antedating any anxiety that includes the word death in its descrip
tion" (Winnicott, 1956, p. 303). This state is so terrifying that it must be
defended against by whatever means are possible, and the only
defense the infant has are omnipotent fantasies. Thus, the infant cre
ates omnipotent fantasies to protect itself from annihilation anxiety
(Winnicott, 1952), and these fantasies cannot be gradually relin
quished in favor of reality. A prisoner of fantasied omnipotence, the
infant protects itself against any incursion of reality into its private
world; indeed, when reality threatens to intrude, it must be reinter
preted according to the child's omnipotent fantasies, or "brought into
the sphere of omnipotence."
This distortion of reality by infantile omnipotence is the essence of
158 Chapter 4
Winnicott's concept of the developmental origins of psychosis. He did

not deny the possibility of constitutional factors in psychosis, but he
believed that impingements play a crucial role by producing annihila
tion anxiety and the omnipotent defenses against it. Fixated in the
omnipotent defense, the ego cannot become integrated, and indeed all
developmental tasks suffer as the child's investment is focused on
protecting itself via omnipotent defenses. "The focus is on the shell,
not the kernel" (Winnicott, 1960a). Since all tension states threaten to
become annihilation anxiety, frustration and disappointment tend to
be denied and "magically" relieved, fixating the personality at the
level of magical thought. Any assault from reality threatens to pierce
the fragile defense and must be reinterpreted in line with the fantasied
omnipotence. Thus, the growing child is driven further from reality
and nearer to the outbreak of psychosis.
These dynamics are illustrated in W innicott's case of Miss X
(Winnicott, 1963d), a schizoid patient who had the habit of covering
herself with a rug in her treatment sessions. In a Tuesday session that
Winnicott reported, Miss. X spent her time covered with the rug and
said very little; although neither she nor Winnicott was anxious, noth
ing much happened. Toward the end of the hour on Wednesday, she
pointed out that the Tuesday session had been crucial because she had
gone through a difficulty that had blocked her previous analysis. That
analyst had not allowed her quietude, and Miss X initially felt that the
Tuesday hour signified the failure of her analysis with Winnicott.
However, during the session she realized, according to Winnicott, that
this analysis "was not going to fail in the usual way, and that she
would go ahead and take all the risks and let feelings develop, and
perhaps suffer deeply. In this way she found this Tuesday hour
extremely satisfying, and she felt grateful" (p. 237). The case of Miss X
demonstrated to Winnicott that what patients like her need is for the
analyst to "know and tell them what they fear. They themselves know
all the time, but the thing is that the analyst must know and say it"
(p. 237). The patient pointed out that she needed Winnicott to "take
over" her omnipotence so that she could "with relief break down,
break up, and experience the worse degree of disintegration or sense
of annihilation" (p. 237). In the Thursday session the patient showed
the first indications of ambivalence in her relationship with Winnicott,
a significant step forward.
Winnicott believed that this vignette illustrated why the analyst
must meet the patient's need for omnipotence in treating borderline
cases. Once the patient felt that she was understood without having to
signal her needs to Winnicott, she was able to "let go" of her omnipo
D. W. Winnicott 159
tence, surrender her minimal sense of self to him, and begin the
process of developing a real sense of self. As the omnipotent defense
was relinquished, the patient experienced the annihilation anxiety that
had been hidden underneath it. In Winnicott's view, Miss X, while not
clinically psychotic, had psychotic needs because of her reliance on
the omnipotent defense to protect against annihilation anxiety.
Winnicott applied a similar analysis to states of "depersonaliza
tion" and "derealization." Both conditions result from impingement in
the phase of absolute dependence (Winnicott, 1962). Winnicott
equated such states with the eruption of an underlying lack of person
alization or underdeveloped sense of reality. In the case of "deperson
alization" episodes, even though a presumably normal or neurotic
personality may have grown around the deficit, the underlying sense
of having a "lived psychic reality" is shaky; there is never a definite
sense of experience as belonging to oneself. Similarly, Winnicott under
stood "derealization" experiences as the surfacing of an underlying
uncertainty about the relationship of the self to reality. Both states are
symptomatic of damaged self development due to impingement in the
phase of absolute dependence, even though the personality may
appear relatively healthy.
The omnipotent defense need not lead to a psychotic personality
organization (Winnicott, 1960a). The result seems to depend on the
severity of the impingement. If the self is completely unintegrated
the likelihood of a psychotic outcome is great, but if the self has
achieved some degree of integration there may be a stable omnipo
tent defense that is in continual conflict with reality but does not
lead to a full-blown psychosis. In these cases conflict and tension are
"brought into the sphere of omnipotence" (Winnicott, 1960a). Bor
derline and narcissistic personality disorders fit Winnicott's concept
of trauma in the phase of absolute dependence with arrest at the
level of omnipotent defense.
Patients at this level have expectations of people and their therapist
that reflect their omnipotent world of absolute dependence. They typi
cally expect the therapist to fulfill all manner of needs and desires. There
is no sense of others, including the therapist, as having lives of their
own, or even feelings of their own. We saw in the case of Miss X that she
needed the analyst to know her fears without being told. Patients'
demands that the therapist bring them home, give them money, reduce
their fees, and hold and caress them, are all, for Winnicott, manifesta
tions of the omnipotence resulting from impingements in the phase of
absolute dependence. Patients expressing such demands are trying
desperately to blur any awareness of others as separate.
160 Chapter 4
Because such patients have no awareness of the other as separate

from the self, they are pre-ruth, unable to discern that their rage can
injure the other and consequently feel no guilt for "injuring" other
people. When such patients react angrily to threats to their omnipo
tence, they can rage, insult, threaten, and even become physically vio
lent without guilt. From the ego psychological viewpoint, this
indifference to hurting others reflects a superego lacuna, a gap in the
structure of the psyche. While Winnicott would not disagree that there
is such a deficit, in his view, this apparent coldheartedness stems from
fixation at the phase of absolute dependence, before the capacity for
concern has a chance to develop.
The therapist's role is to attempt within the therapeutic context to
meet as much as possible the patient's previously unmet need for life
within the sphere of omnipotence (Winnicott, 1954c). The therapist
attempts to understand the patient's behavior as communication even
when the patient does not know he or she is communicating. To
achieve this goal, the therapist uses the empathy of the "good enough
mother," who relies on identification with the baby to meet uncom
municated needs. Because psychopathology at this stage is due to pre
mature interference in the delusional omnipotent world of absolute
dependence, Winnicott (1954c, 1960a) believed that the therapist
should offer the patient, insofar as possible, the opportunity to return
to this world. Such patients need to feel what they should have felt as
a baby: dependence without awareness, the omnipotent feeling that
needs are met magically without communicating them. The therapist's
sensitivity to this need and effort to meet it, when possible, return the
patient to the point of ego arrest, from which the maturational process
can be unblocked.
This principle is illustrated by Winnicott's (1954b) discussion of a
schizoid-depressed patient who was not generally regressed but who
lapsed into momentary withdrawal states. In one session the patient
had the fantasy of curling up and rolling over the back of the couch. A
few weeks later when he asked Winnicott for advice and Winnicott
did not comply, the patient withdrew and felt the desire to curl up.
Winnicott interpreted that the patient's throwing himself around and
curling up implied the "existence of a medium" (p. 256). Soon after,
the patient had a dream of discarding a shield, which Winnicott inter
preted as a symbol of his unneeded defense now that he (Winnicott)
had proven himself a suitable medium for the patient at the moment
of regression: "It appears that through my immediately putting a medium
around his withdrawn self I had converted his withdrawal into a regression,
and so had enabled him to use this experience constructively" (p. 257).
D. W. Winnicott 161
Understanding the patient's need for a lap to curl up on, Winnicott

provided an "analytic lap." Considerable therapeutic progress fol
lowed these events. The patient gained a clear understanding of the
analyst's role, resulting in his management of reality situations at
home and work in a new way. Additionally, he began to feel concern
for his mother.
This vignette illustrates Winnicott's idea that patients who are
not generally regressed often need moments of regression to return
to prim itive parts of the self that are fixated in the phase of
absolute dependence. This patient needed to return to dependence
without awareness in order to advance to ambivalence and the
confrontation of reality situations in his life. By understanding his
need for a "m edium," a "lap" to curl up in, Winnicott met his
need for a brief return to absolute dependence in order to allow
movement into relative dependence.
This approach highlights Winnicott's view of psychotherapy as
adaptation to need. Winnicott (1959) made clear that he was giving
up his traditional view of psychotherapy as the making of "clever"
interpretations in favor of a view of psychotherapy as the analyst's
adaptation to the patient's unmet needs. The therapist's role is to
assess the point of developmental arrest in each patient and then to
make an adaptation to the arrested need. Verbalization is only one
way this may happen, but for preoedipal disorders words are often
not suited to the task.
A good example of Winnicott's principle of adaptation to regres
sive needs in a functional patient is the case of a 47-year-old woman
who had made a seemingly positive adjustment to the world, was
earning a living, and was well liked (W innicott, 1949, 1954b).
Despite these outward signs of health, she felt completely dissatis
fied with her life, having never felt in contact with herself. A classi
cal analysis had not affected this core belief. W innicott (1949)
described her treatment as follows:
With me it soon became apparent that this patient must make a very severe
regression or else give up the struggle. I therefore followed the regressive
tendency, letting it take the patient wherever it led; eventually the regres
sion reached the limit of the patient's need, and since then there has been
a natural progression with the true self instead of a false self in action-----
In the course of the two years of analysis with me the patient has repeatedly
regressed to an early stage which was certainly prenatal [p. 249].
In this patient's previous analysis she had hysterically thrown

162 Chapter 4
herself off the couch. Winnicott interpreted this behavior as her

unconscious need to relive the birth process. In her treatment with
Winnicott she relived the birth experience over and over again. The
patient experienced the anxiety of having her head crushed; she man
aged this anxiety by identifying with the crushing mechanism, obtain
ing gratification from the fantasy of destroying her head, which no
longer felt like part of the self. The patient "had to accept annihila
tion," which was experienced as a giving in, and "eventually the
appropriate word was 'a not knowing' "(p. 250). The acceptance of
"not knowing" gave the patient tremendous relief and was, as Winni
cott put it, "transformed into the analyst knows, that is to say, behaves
reliably in active adaptation to the patient's needs" (p. 250).
In Winnicott's view, this case illustrates a functional patient's need
for regression to absolute dependence and the analyst's role of making
an "active adaptation" to this need. W innicott understood the
patient's throwing herself off the couch as a need to relive the birth
process, but he did not believe the treatment could progress solely by
this understanding. He used his formulation to help the patient relive
the birth process and, thereby, to experience the annihilation she had
been defending against with her superficial adaptation to reality. The
key to the therapeutic process was, according to Winnicott, her ulti
mate willingness to "give herself over" to the analyst—that is to say,
to accept absolute dependence—so that she could begin to form a true
sense of self. Thus, Winnicott conceptualized the analyst's role as one
of managing the patient's dependency needs rather than providing
interpretations; the analyst uses his understanding to adapt to the
patient's need to regress to absolute dependence and annihilation,
from which point the patient is able to move forward.
Winnicott's use of general management in severely regressed
patients is also illustrated in his treatment of the analyst Margaret
Little. In her description of her analysis with Winnicott, Little (1985)
pointed out that Winnicott made few interpretations, preferring to
manage her regression primarily in nonverbal fashion. In the first ses
sion Little lay silent, curled up tight under a blanket; Winnicott said
nothing until the end of the hour, when he commented, "I don't
know, but I have the feeling that you are shutting me out for some
reason" (p. 20). Some weeks later Little was seized with spasms of ter
ror, and Winnicott, who believed she was reliving her birth and said
as much, let her cling to him until they passed. In many sessions Little
lay inert under the blanket and Winnicott held her hands tightly in
his. Little described Winnicott as generally "holding" her, taking full
responsibility, including management of her life when necessary. For
D. W. Winnicott 163
example, once he was concerned about her ability to withstand his

vacation, so he arranged, without her knowing, for a friend to invite
her to Switzerland. On another occasion, he hospitalized her while he
was away, and during a period when she could not reach her analytic
sessions he came to her house.
It is clear from Little's description of her analysis with Winnicott
that he attempted to meet the needs of her state of absolute depen
dence by overall management of her care. He did make interpreta
tions, but they were not as important as his adaptation to her
regressed state. Winnicott attempted to function like the "good
enough" mother of an infant in the phase of absolute dependence. As
Little began to do more for herself, he "de-adapted" just as the mother
does when the infant reaches the phase of relative dependence.
Winnicott (1963c) believed interpretations were effective only
when they adapted to the patient's needs. He reported an example
of such an interpretation in the treatment of a woman who had
recently changed her analytic schedule from once to five times per
week when Winnicott announced his impending month-long vaca
tion. Before the announcement she had a dream in which a tortoise
with a soft shell was killed to relieve its misery; the dream indicated
to Winnicott that she had come to treatment for relief of her suicidal
tendency. After the vacation announcement the patient became
physically ill, and Winnicott linked the illness with his impending
departure. The vacation reenacted traumatic events of the patient's
childhood. Winnicott wrote, "[The patient felt] as if I were holding
her and then became preoccupied with some other matter so that
she felt annihilated" (p. 250). Winnicott then interpreted to the
patient that by killing herself she would gain control over the anni
hilation associated with feeling dependent and vulnerable. Further,
he pointed out that the illness was a physical localization of her
urge to die. This interpretation relieved the patient's feeling of help
lessness, and she was able to allow Winnicott to go on vacation.
Winnicott described his understanding of the effectiveness of his
interpretation as follows:
One can only assume that understanding in a deep way and interpret
ing at the right moment is a form of reliable adaptation. In this case, for
instance, the patient became able to cope with my absence because she
felt (at one level) that she was now not being annihilated, but in a posi
tive way was being kept in existence by having a reality as the object of
my concern. A little later on, in more complete dependence, the verbal
interpretation will not be enough, or may be dispensed with [p. 250].
164 Chapter 4
Winnicott's interpretation turned a potentially traumatic repetition

into a new, positive experience. The value of interpretation, for Winni
cott, lies in the meeting of a childhood need, in this case, the need to
be recognized within a dependent relationship. (As we will see in
chapter 6, Kohut too held that the effectiveness of interpretations lies
in the type of experience they provide, rather than in the accretion of
knowledge, although his conceptualization of their impact differs
from Winnicott's.)
Winnicott did not believe that the therapist will always be suc
cessful in understanding and meeting the patient's need. The thera
pist's failures are inevitable, just as the mother's adaptation to her
child is never perfect.
Just as the mother's failures help the infant achieve a sense of real
ity by forcing the infant out of the delusional world of omnipotence,
so too do the therapist's failures in adaptation force the patient to see,
however briefly, that the therapist is not under his or her omnipotent
control (Winnicott, 1963b). The patient's recognition of the therapist's
limitations is equivalent to the mother's "bringing reality to the child."
It is as important in the former case as in the latter that too large a
"dose" of reality not be given too soon. The therapeutic process of
adaptation and failure, if not so severe as to be traumatic, aids in the
development of the patient's sense of time and the continuity of expe
rience. Like the mother, the therapist bears the greatest burden of syn
chrony with needs deriving from the absolute dependence phase. Just
as failures in environmental adaptation become more bearable and
can be born for a longer period once the baby develops a temporal
sense, so too is there more leeway for therapist error with later devel
opmental issues. In short, the therapist's role is the same as the moth
er's: to meet the needs as well as possible and then gradually bring the
patient/child into reality by a process of adaptation, failure, and
repair that allows reality to be experienced in manageable doses.
Critical to the success of the therapeutic process is the therapist's
willingness to acknowledge "failure." Winnicott felt that an exclusive
reliance on interpretation of the patient's reaction to the therapist's
failures does a disservice to the patient because it does not use the
mistake to help the patient achieve a sense of reality. Any patient dis
rupted by the therapist's error must have at least some residual fan-
tasied omnipotence not fully brought into the reality sense; for this
reason, acknowledgment of failure is a therapeutic necessity. The ther
apist's willingness to admit error acknowledges his or her limits and
brings to the patient the reality of the therapist as a person, thus, for
that moment, breaking through infantile fantasized omnipotence. Just
D. W. Winnicott 165
as the infant gradually learns reality through failure and frustration,

so too must the patient be gradually "brought to reality" by the
failures of the therapist and their realistic acknowledgment.
Winnicott (1951) also believed that the origins of fetishism were to
be found in the absolute dependence phase. Like Freud, he felt that
such attachments to objects were encapsulated psychoses, based on
denial of reality. Whereas Freud believed the basis of the fetishistic
attachment was a sexual denial, Winnicott believed the psychotic
attachment had a presexual origin. Unlike the child, who knows that
the transitional object is not the mother, the fetishist believes the
fetishistic object to be the longed-for object. According to Winnicott,
the fetish originates in an impingement in this phase, which forces the
infant out of its delusional world and into reality before the ego is
ready. To defend against this premature awareness of reality, the
fetishistic object is brought into the "sphere of omnipotence" where
the magical wish for the need to bring its own gratification is satisfied.
Such an object is not used as a transitional object is, that is, to shift
from om nipotent control to lim ited control, but enables the
child/patient to stay within the orbit of omnipotent control; because it
must always be available to satisfy magical desires it cannot be given
up. Whereas the transitional object is known to be outside omnipotent
control but is treated as though it were not, the fetishistic object is not
experienced as outside. It is "pretransitional," and therefore the
attachment to it is delusional as opposed to illusional.
The treatment for fetishism follows the same principles as ther
apy for any symptom originating in the absolute dependence phase.
The therapist's role is not to try to teach the patient reality by inter
preting the attachment to the object as unrealistic. Even if the inter
pretations are thoughtful, in-depth, and accurate, words cannot
alter omnipotent defenses needed to protect against annihilation
anxiety; words have little meaning to the patient who needs to be in
the orbit of infantile omnipotence.
The most benign outcome of impingement in the phase of absolute
dependence is "ego distortion in terms of true and false self" (Winni
cott, 1960b). If impingement is not too severe and some integration of
self has occurred, the child may be able to defend itself by environ
mental compliance. That is, the child attempts to provide the environ
ment with what seems to be expected. In order to protect the minimal
sense of "true self," the environment is offered a personality. The
defense protects the buried "true self" from contact with the environ
ment and in so doing forfeits feelings, the very sense of aliveness. The
surface personality looks normal and is taken by the world as the real
166 Chapter 4
personality. However, it does not feel real to the patient, whose

principal feelings are boredom, restlessness, and emptiness. This
malaise results from a lack of connection to inner experience.
"Authenticity" is a crucial category for Winnicott, designating the
healthy personality governed by the true self. Living inauthentically
is a defensive reaction, a distorting of the ego to protect all "real"
feelings from environmental contact. Unlike cases in which the
unintegrated personality is predisposed to psychosis, in cases of ego
distortion a self exists, even if only in a nascent state, but its buried
under a constructed "normal" personality.
We have already seen an example of this split between the true and
false selves in the case of the 47-year-old woman who seemed to have
effected a normal adaptation but felt disconnected from herself. The
patient's personality and life had been built around mental function
ing; she lived in her head and was disconnected with the rest of her
self. Consequently, she had no sense of self until Winnicott followed
her regressive need to absolute dependence, entailing fear of annihila
tion, the "limit of her need." When the patient reached this point of
the treatment, she reached the buried true self, and for the first time
felt connected with herself.
Winnicott felt that such constructed false selves are more com
mon than is usually recognized even by psychoanalysts and that
they tend to be analyzed as though they were real. He mentioned
cases of patients previously analyzed along classical lines where
the patients' "nonexistence" had to be recognized before therapeu
tic movement could begin (Winnicott, 1954c, 1960b). To analyze
wishes, conflicts, and oedipal issues in such cases is meaningless
because the patient does not feel real. The patient only feels commu
nicated with when his or her fundamental nonexistence is recog
nized; only at that point can the analysis have a m eaningful
beginning. Winnicott (1960b) reported a case of a male patient who
had had an extensive previous analysis.
My work really started with him when I made it clear to him that I recog
nized his non-existence. He made the remark that over the years all the
good work done with him had been futile because it had been done on
the basis that he existed, whereas he had only existed falsely. When I had
said that I recognized his non-existence he felt that he had been commu
nicated with for the first time. What he meant was that his True Self that
had been hidden away from infancy had now been in communication
with his analyst in the only way which was not dangerous. This is typical
of the way in which this concept affects psycho-analytic work [p. 151].
D. W. Winnicott 167
The very feeling of unreality is the analytic issue, and the analyst's
task is to point out the unreality of the false self until the true self can
begin to emerge. Insofar as it helps to get beneath the false self,
defense interpretation is useful, but when it treats the false self as
though it were real, it is destructive to treatment. The analyst's task is
to make contact with the true self, as Winnicott did in this case by
pointing out the patient's nonexistence.
Psychopathology and Relative Dependence
We have seen that as children move into relative dependence by

becoming aware of the mother as a separate person, they begin to
internalize the mother and develop a sense of self as continuous and
integrated. Impingement in this phase occurs to a developing self—not
to an unintegrated self, as in absolute dependence, nor to a fully inte
grated self, as in the oedipal phase. Having achieved some degree of
internalization, the infant experiences impingement as deprivation, a
loss of something that was there, as opposed to the "privation" of
impingement in absolute dependence.
According to Winnicott, if mothering has been good enough until
some point in the phase of relative dependence and then fails signifi
cantly without repair, the child will seek to regain what has been lost.
Efforts to regain the lost object gain expression in the various symptoms
of character disorders and borderline cases. One typical reaction is to
attempt to replace maternal deprivation by taking objects. Winnicott
(1956b) understood child and adolescent stealing as efforts to regain
the lost mother by laying claim to the world. Children attempt to steal
back what was stolen from them. In Winnicott's view, such behavior
reflects a belief based on an unconscious memory of gratification that
there is something of value to be found. Because an attempt to find the
object is really an effort to "re-find" it, reflecting deprivation of
something once possessed, Winnicott considered stealing a positive
indicator.
If the environment understands the effort to re-find in the act of
stealing, the behavior can be halted. For example, Winnicott (1963d)
reported the case of an 8-year-old boy who had been caught stealing.
When Winnicott interpreted to him that he was searching for the
"good mother" from the time before she conceived his younger sib
ling, the delinquency ceased. In some cases the treatment must go
beyond interpretation by meeting dependency needs antedating the
deprivation. For example, a 13-year-old boy was not only slashing
sheets but also stealing at school and getting into other serious
168 Chapter 4
trouble. Winnicott arranged to have him "mentally nursed" at home

while he underwent a period of regression from which he sponta
neously emerged. The boy ceased his delinquency and eventually
returned to school. In Winnicott's view, the "mental nursing" pro
vided the boy with the "good mother" for whom he had been search
ing, thus obviating his need to steal. In both cases the therapeutic
environment adapted to the patient's need for the gratification that
was once obtained and then lost. When this need was met, the search
for the lost object ceased.
Some patients attempt to re-find the lost object through direct
physical gratification, which leads to addictions involving food,
drugs, and alcohol as well as to sexual promiscuity (Winnicott, 1951).
Winnicott believed that such addictions are similar to transitional phe
nomena because they represent the mother and yet are recognized as
not being the mother. Unlike the child's transitional possessions,
addictions are not given up naturally because they are a regressive
response to deprivation, an effort to regain an earlier relationship with
the mother before deprivation. Consequently, rather than moving the
developmental process forward, addictions tend to remain fixated.
Some babies suffer impingement in the phase of relative depen
dence not from an object taken away prematurely, but from a mater
nal environment unable to shift from the "near-absolute" adaptation
of absolute dependence to the lesser degree of adaptation necessary
for the next phase (Winnicott, 1960a). In such an environment the
mother-infant "set-up" continues to be based on object instinctual
excitement rather than fostering the development of an "ego relation
ship." The growing infant has a continual need for contact, to know it
is in a relationship, and thus lacks the capacity to be alone (Winnicott,
1958). While the infant is differentiated enough to have a relationship,
it is not able to maintain a sense of the other without physical pres
ence. The infant has enough of a "lived psychic reality" to feel alive
and whole when in the presence of the mother but not enough to feel
alive and whole in her absence. Patients fixated at this level cannot be
alone without feeling lonely and desperate for physical contact to
regain the feeling of life and reality. Such patients become addicted to
others, either through excessive dependence or a need for sexual con
tact; that is, they use people for the same purpose other patients use
food, drugs, or alcohol.
Because the inability to be alone is created when the child is not
allowed "to be alone" in the mother's presence, patients who cannot
be alone need to have the experience with the therapist they should
have had with the mother: they need to be alone in the presence of the
D. W. Winnicott 169
therapist (Winnicott, 1958). If the therapist insists on verbal interven

tion, the patient is denied this critical therapeutic experience. The ther
apist's role is to be a presence without intruding on the patient's need
to be alone—which will usually be manifested when the patient
chooses to be silent. Silence in such patients is not to be interpreted as
resistance but as the patient's effort to unblock the effects of not hav
ing been allowed to be alone and resume self-development by the for
mation of an ego relationship. The therapist's silence offers the patient
a sense of connection without physical contact so that the patient can
internalize the therapist as a good object. Once the therapist is inter
nalized, the patient is able to be alone without the anxiety of losing a
sense of existence.
Winnicott's (1971) treatment of the patient who lacks the capacity
to be alone is illustrated by his rather detailed description of a session
with a patient who found, after a six-year analysis of five sessions per
week, that she needed sessions of indefinite length. To accommodate
this need, Winnicott saw her once a week for three hours, later
reduced to two hours. In the session described by Winnicott, the
patient said very little for the first half hour during which she either
sat on a chair or on the floor or walked around the office. She then
commented that she was unable to be herself, that she was "not really
looking," and that there was "just a mess inside, just a crash" (p. 58).
After about one hour, with her head on a cushion, she mentioned a
dream of a girl bringing pictures that showed no improvement; once
again she commented that she did not feel a "me." She was mostly
silent for the next half hour and then said, "Just drifting like clouds"
(p. 59). She began to talk about being of no consequence and of
attempting to feel that she mattered by conforming to what was
expected of her. She then mentioned with surprise that a girl sent her
a postcard. Winnicott commented, "As if you mattered to her. But you
don't matter to her or anyone" (p. 59). The patient commented that
she had not made contact with Winnicott all day.
Winnicott noted that after two hours the patient for the first time
"seemed to be in the room with [him]" (p. 60). It was a makeup ses
sion and she said she was glad Winnicott knew she needed the time.
The remark was made as though it were her first comment to Winni
cott that day. The patient began to look back on the session but could
not remember what she had said. At this point, Winnicott made his
first interpretation: "All sorts of things happen and they wither. This
is the myriad deaths you have died. But if someone is there, someone
who can give you back what has happened, then the details dealt with
in this way become part of you, and do not die" (p. 61). When a little
170 Chapter 4
later the patient expressed the feeling that she had come to meet
someone who did not appear, Winnicott began to make links between
her comments and to reflect back what she was saying. At one point,
the patient realized that she had interrupted Winnicott, indicating her
acceptance of his existence. They went on to discuss how talking to
herself did not help her feel a sense of existence unless it was a contin
uation of talk reflected back to her by another. The patient com
mented, "I've been trying to show you me being alone” She talked of
her use of mirrors to search for herself, and Winnicott pointed out that
it was she who was searching. Winnicott reported that by the next ses
sion the patient had forgotten this session and they took two hours to
reach this same point again. In the second session the patient asked a
question and said with deep feeling that "one could postulate a ME
from the question, as from the searching" (p. 64).
This session clearly reveals Winnicott's technique, but the impor
tant point in the present context is his willingness to allow the
patient to be alone in his presence. He made no effort to get her to
talk and asked no questions. Allowing her to be alone until she made
contact with him. By beginning to "come into the room" and recog
nize his existence, she began to form an ego relationship, thus initiat
ing the development of the capacity to be alone, a crucial step in her
self development. The fact that this patient could not take this step
without Winnicott's presence and attention to her behavior illus
trates a crucial point in Winnicott's theory of technique. Winnicott's
task was to "hold" the situation by linking the various moments of
the session and giving them back to her. He performed the function
of self integration by giving her self back to her in coherent form. It
was Winnicott's belief that he was performing the task of ego organi
zation, which the parents had failed to perform in the patient's child
hood. This belief was borne out in the next session when the patient
recognized herself as existing in her question.
It is not that Winnicott believed that silence could not be resistance
as it is viewed in classical psychoanalytic theory. Rather, he believed
that it has different meanings depending on the developmental level
of the patient and the stage of the treatment process. In a neurotic
patient anxious about intense affects, silence is a resistance. For the
preoedipal patient trying to develop the capacity to be alone, silence is
a developmental advance toward the creation of the self via an ego
relationship. Illumination of the therapist's responsibility to adapt to
this need is one more way in which Winnicott that felt his studies of
preoedipal disorders expanded psychoanalytic theory and technique
without questioning its relevance to neurotic disorders.
D. W. Winnicott 171
Confusion results because neurotic patients, too, may have some

arrest in the phase of relative, or even absolute, dependence.
Although these issues would be more transitory in such cases, they
may still require intervention geared to the preoedipal disturbance
at those "regressive moments." If one such moment involves an
unresolved need to be alone, silence may emerge with the uncon
scious purpose of affording the patient the experience of aloneness
with a mother who will tolerate being ignored. In this situation,
requests for free association or interpretations of resistance will
interfere with the patient's developmental need to be alone and will
reenact the behavior of the mother who intruded with contact and
stifled the child's need to form an ego relationship. Whether verbal
intervention is used in any individual situation depends on the
assessment of the patient's developmental level.
Many borderline personality disorders fit Winnicott's (1960,1963a)
concept of trauma in the phase of relative dependence. Winnicott
viewed the overwhelming dependence ties and acting out so charac
teristic of such patients as efforts to "lay claim" to something in the
world, to use people or things as transitional objects in an effort to
regain a lost maternal bond. Borderline patients often create burden
some countertransference strain because of the demands and expecta
tions they place on the therapist. Common clinical terms used to
describe this phenomenon include "entitlement," "lack of bound
aries," and "unrealistic expectations." As we shall see when we con
sider Kernberg's work in chapter 5, treatment recommendations for
such patients often involve limit setting and confrontation with reality
(Kernberg, 1975). Masterson (1976) also subscribes to this approach.
In Winnicott's view, all such recommendations ignore the develop
mental meaning of the patient's behavior. He believed that since the
patient's expectations and demands are an effort to lay claim to an
object once possessed but now lost the therapeutic task is to adapt as
much as possible to the patient's longing for a gratifying object. By so
doing, the therapist places himself in the position of a transitional
object—to be used as a symbol for maternal gratification and dis
carded when no longer needed. According to Winnicott, this type of
adaptation unblocks the arrested maturational process.
The case of Miss X discussed earlier illustrates the meaning of the
borderline patient's "claim" to others (Winnicott, 1963e). Recall that in
one session Miss X regressed to a desire to have Winnicott know her
needs without any communication from her, thus reflecting her
omnipotence. In the next session, she became greedy and expressed
the desire to "eat" Winnicott. They discussed the "com pulsive
172 Chapter 4
element" in her appetite and its manifestation in her relationship

with Winnicott, and it became clear that there was a "compulsive
greed in her antisocial tendency" (p. 238). The meeting of her
regressed need from the phase of absolute dependence had clearly
led to a movement toward relative dependence. She now recognized
Winnicott as an object upon whom she was dependent and claimed
him as her own. This movement was also manifested in a new
ambivalence toward Winnicott.
The borderline patient, according to Winnicott (1960a), lives in the
"intermediate area" of infancy when the child goes back and forth
between merger and separation. During merger, the mother/therapist
is expected to know the infant's/patient's needs without being told
even at those times when the infant/patient has a need to feel sepa
rate; yet such maternal empathy is also felt as a stifling assault on the
need to separate. Therefore, the patient's "claims" on the therapist are
contradictory; moreover, they oscillate quickly and contain no signal
about which need is being expressed at any given moment. The thera
pist inevitably feels intruded upon, confused, dejected that he or she
cannot do anything right, and, ultimately, enraged at the patient. The
twofold therapeutic task is to understand that the ricocheting behav
iors are symptomatic of a developmental process arrested in the tran
sitional phase and to meet the need for a transitional object to unblock
the arrested maturational process. Limit setting from this point of
view misses the patient's transitional level of development. If such
patients respond to limit setting, they are only conforming to the
expectations of the therapist.
Such conformity can have no emotional meaning to them and their
need for the lost object can only be buried. Limit setting not only
means lost therapeutic opportunity, according to Winnicott, but is
antitherapeutic in that the patient's desperate effort to signal for help
is quashed. The therapist must present himself or herself as an object
to be merged with and separated from, as needed, just as the transi
tional object must be available for comforting when it is needed, only
to be forgotten about at other times.
Insofar as interpretation is useful for borderline patients, it tends to
emphasize the patient's unconscious object seeking and the way the
patient needs to use objects. However, Winnicott seemed not to rely
heavily on interpretation in borderline and other preoedipal cases.
Rather, he stressed the importance of "using" the therapist as an
object. (In Winnicott's, 1971, writings this expression refers to the way
an object is employed to aid the developmental process; it does not
imply an exploitative motive.) As we saw, in the phase of absolute
D. W. Winnicott 173
dependence there is no use of an object because there is no awareness

of objects. It is only when the child "expels" objects out of its sphere of
omnipotence, beginning with transitional possessions, that it learns to
"use objects." To be used as a transitional object the therapist must
withstand an intense variety of feelings, be there when needed and
forgotten when not needed.
For the patient to use the therapist in this way the therapeutic space
must be formless so that the patient can create meaning out of it. The
therapist must be careful to avoid imposing preconceived categories
on the patient so that the patient can regress to a blank space and cre
ate a new object endowed with personally created meaning, just as the
child creates its transitional object. If the therapist does not interfere
by making intrusive, premature interpretations or otherwise imposing
meanings on the therapeutic space, the patient is able to use it to cre
ate his or her own meaning. This process is crucial to shifting the ana
lytic process from interpretation of the past to creation of new
meaning. According to Winnicott, creativity plays a major role in all
effective analyses because only through creativity can one give one's
own meaning to reality, that is, enter the process that forms the self. In
normal development the transitional object is used for this purpose; in
analytic treatment the analytic space functions in an analogous way
by facilitating creation of personal meaning. This principle holds true
to some degree for all analytic treatment, but it is especially significant
for the borderline patient, whose emotional development is arrested at
the transitional phase. For such a patient there is no internalized
maternal image; the primary transference issue is this absence rather
than the repetition of an existing maternal image, as seen in higher
level disorders. Consequently, the creation of a new object in the
absent space is a major outcome of successful treatment.
The creation of meaning out of formlessness is illustrated in the
three-hour analytic session described earlier. Winnicott made no
effort to get the patient to talk or free-associate. He allowed her to use
the session as she saw fit. This atmosphere is what Winnicott means
by the "formlessness" from which the patient is able to create per
sonal meaning. In the session in question, the patient eventually cre
ated a sense of herself and Winnicott that felt real and meaningful to
her. Recall also that this process was made possible by Winnicott's
reflecting back "bits" of the patient in coherent form. This mirroring
function of the analyst makes the creation of self possible as the
patient moves from the sense of being an absence to that of being a
newly created presence.
This movement from absence to presence is also illustrated by
174 Chapter 4
Winnicott's (1971) treatment of a middle-aged woman who tended to

lapse into dissociated states. The patient felt that nothing had signifi
cance for her, that she did not exist in her own right, and that all her
activities were meaningless. In the analysis she realized that no one in
her childhood had recognized her need for formlessness and that she
had been molded into forms for others, none of which had meaning to
her. As she reached this realization, she became intensely angry. At
the next session she expressed fear that Winnicott would interpret
recent constructive changes in her life as evidence of therapeutic
progress. For example, in describing how she had cleaned up messes
that had been neglected for months, she expressed the suspicion that
Winnicott had somehow prearranged this change. Winnicott pointed
out to her that he had to be careful not to show any pleasure in her
activities lest she feel he had fit her into a pattern. The patient asked
what the previous session had been about, and Winnicott described it
as formless. She said that she was feeling tired and realized she could
sleep in the session. This thought allowed her to imagine being
healthy, and she had the idea that she "might be able to be in charge
of [herself]" (p. 36). Eventually she arrived at the conclusion that
in illness she was certain of what to expect but in health she had the
anxiety of uncertainty. Winnicott felt that a great deal had been
accomplished, but he did not allow himself to become pleased.
This case demonstrates Winnicott's ability to let the therapeutic
space be used as material for the creation of a transitional object. The
patient's anger had a profound impact on her because it sponta
neously erupted out of the formlessness of the therapeutic space. Win
nicott believed that if he had interpreted her anger toward childhood
figures, the interpretation, although correct, would not have belonged
to her and therefore would have impinged on her need to possess her
experiences as her own. In fact, he referred to such premature inter
pretations as "stealing" from the patient. One of Winnicott's cardinal
therapeutic principles is that therapeutic effectiveness is greatly
enhanced if the patients themselves make the interpretation, because
in the process they take a step in the creation of the self.
Winnicott (1971) summarized the use of therapeutic space for the
creation of the self this way:
The searching can come only from desultory formless functioning, or

perhaps from rudimentary playing, as if in a neutral zone. It is only
here, in this unintegrated state of the personality, that that which we
described as creative can appear. This if reflected back, but only if
reflected back, becomes part of the organized individual personality, and
D. W. Winnicott 175
eventually this in summation makes the individual to be, to be found; and

eventually enables himself or herself to postulate the existence of the self.
This gives us our indication for therapeutic procedure— to afford
opportunity for formless experience, and for creative impulses, motor
and sensory, which are the stuff of playing. And on the basis of playing
is built the whole of man's experiential existence [p. 64].
The patient's need to use the therapist as a transitional object makes

the countertransference an especially difficult issue with such
patients. Winnicott (1960c) defined countertransference as disturbance
in the therapist's observational stance. As the patient demands com
pliance with expectations that the therapist function as a transitional
object, the therapist experiences a countertransference strain. Such
patients are demanding that the therapist yield the professional
analytic attitude, and they become enraged whenever they feel
the therapist has not complied with their expectation. This strain
makes working with such patients more emotionally draining than
the psychotherapy of neurotic patients, and a major component of the
therapeutic work is the management of these very difficult counter
transference feelings. According to Winnicott, countertransference
feelings—which range from anger, disruption, and feeling threatened
to love and affection—must be felt by the therapist. The danger lies
not in having countertransference feelings but in repressing them, for
repression can lead to countertransference acting out. The contain
ment of these intense feelings imposes a burden on the therapist as
heavy as the mother's when she manages her infant's distress. Fur
ther, since the best therapeutic approach is often to remain silent, the
therapist is denied even the outlet of interpretation. The therapist's
absorption and management of personal feelings must be outside the
patient's awareness, just as the mother bears the strain of nurturing
the baby outside its awareness.
According to Winnicott (1960c), it is an inevitable part of the treat
ment of borderline patients that they will "break through the barriers
that [are] called the analyst's technique and professional attitude, and
force a direct relationship of a primitive kind, even to the extent of
merging" (p. 164). Borderline patients require and even force the
needed primitive relationship, and therapists have no choice but to
alter their professional attitude. Winnicott believed that in most cases
patient and therapist will develop the needed relationship in an
orderly way and will "recover from it" in a similar fashion. Adapta
tion to the borderline patient's need for a regressive relationship is the
crux of the therapeutic effort for such a patient.
176 Chapter 4
A special countertransference problem is created by patients who

need the therapist to hate them (Winnicott, 1947). According to Winni
cott, in psychotic and borderline cases the therapist will inevitably
"hate" the patient at some point during the treatment process. This
feeling may also occur in the analysis of neurotics, but hate can be
kept latent in those cases much more easily. Thorough awareness of
hate is required in the treatment of some severely disturbed patients
because the therapist's hatred is a crucial aspect of the patient's
pathology. For example, Winnicott (1947) found an obsessional
patient loathsome for years. When the patient became lovable, Winni
cott realized that his unlikableness had been a symptom. It was a
major advance in his treatment when Winnicott was able to tell him
that he had hated him but that he had been too ill to be told. Winnicott
referred to this type of hatred as "objective"; he believed that the ana
lyst is not only justified in feeling it but also that he must feel it in
order to treat the patient effectively.
Winnicott believed that the mother hates the baby no matter how
deep her love, long before the baby is able to hate her. As we have
seen, the growing child has a developmental need to integrate love
and hate (Winnicott, 1963b); the infant must learn to hate the mother
without the feeling of destruction. For Winnicott, infants learn to hate
their mother from the mother's hatred of them; it follows that primi
tive patients can only learn to hate if the therapist first hates them.
Much of borderline patients' abusive behavior is designed uncon
sciously to evoke the therapist's hate so that they can experience their
own hatred of the therapist. Patients are then in a position to integrate
love and hate of the therapist/mother.
The inability to integrate aggressiveness and love is a problem
found in a variety of character-disordered patients (Winnicott, 1963a).
Particular forms of impingement in the phase of relative dependence
may have interfered with the child's ability to integrate aggressive
ness with love for the mother. If the mother does not survive, the child
is left with the unconscious interpretation that its aggression killed the
mother. In this situation, unresolved guilt over aggressiveness is
inevitable and will result in one of several forms of defensive reaction.
More commonly, the mother may physically survive but be unable to
contain the child's aggression. If she is made unduly anxious by it, the
child will become anxious over its own aggressiveness and will dis
avow the aggressive component of its personality. The child is then
unable to fuse its love and anger, and this "defusion of the instincts"
splits off aggressiveness from the rest of the personality.
For Winnicott, full object instinctual gratification and excitement
D. W. Winnicott
require the integration of aggression, a "fusion" of the two impulses.

In addition, because the constructive ability to pursue realistic goals
is based on the "aggressive impulse," the absence of aggression blocks
the achievement of these goals. Regardless of the specific pathological
reaction that results from the inhibition of aggressiveness, its inhibi
tion means that a primary source of motivation and zest for life has
been squeezed out of the personality, resulting in its weakening and a
major loss of opportunity for fulfillment.
Various defensive maneuvers may be used to wall off aggression,
all of which rob the personality of its aggressive component. Per
haps the most widely discussed defense against aggression is its
splitting from the erotic impulse (e.g., Grotstein, 1986). As we will
see in the next chapter, Kernberg (1975) views this defensive
maneuver as a pathognomonic feature of the borderline personality
disorder. For Winnicott, splitting is characteristic of any condition
in which depressive anxiety causes aggression to be kept separate
from objects of dependence and love. Patients who must maintain
this split are forced to deploy aggression toward other objects, lead
ing to a continual search for "bad objects" toward whom aggression
can be directed. Conversely, dependent relationships must be kept
conflict free, with any disturbance in them being experienced as a
threat. This constellation of object relationships means that enemies
must always be found and maintained and that love relationships
will lack intensity and full excitement, since all aggression has been
removed from the erotic impulse.
A more benign defense against aggression is reaction formation,
whereby the child turns all aggressiveness into its opposite, is unable
to experience aggression toward an object, and is unable to deploy
aggression in the pursuit of life goals or ambitions. Since any conflict
with others must be denied, the patient denies a great deal of reality to
keep all relationships consistently positive. Further, when an aggres
sive response is needed for self-protection, the individual, having a
general character inhibition, is unable to provide it. This formulation
applies to patients typically described as passive or inadequate. It was
Winnicott's (1950) belief that if aggressiveness is not integrated into
the personality in the phase of relative dependence, the constructive
goal-related pursuits of the individual are severely compromised and
the result is passivity and inadequacy.
This principle is exemplified by a patient who was given to fre
quent destructive outbursts (Winnicott, 1963b). He began an analytic
hour by expressing the wish that his analysis would be of value to the
world. During this session the patient achieved a new awareness of
178 Chapter 4
his envy of Winnicott for being a good analyst; he also wished to

thank Winnicott for meeting his needs.When Winnicott linked his
destructive urge with his view of Winnicott as a good analyst the
patient agreed but was relieved that Winnicott had not interpreted his
wish to contribute to the world through his analysis as a defense
against his wish to destroy in response to the initial remark. Winnicott
(1963b) explained:
He had to reach to the destructive urge before I acknowledged the repa

ration, and he had to reach it in his own time and in his own way. No
doubt it was his capacity to have an idea of ultimately contributing that
was making it possible for him to get into more intimate contact with
his destructiveness. But constructive effort is false and meaningless
unless, as he said, one has first reached to the destruction [pp. 80-81].
In Winnicott's view, the destructive and constructive urges are

inextricably connected. Patients who cannot "reach to" their aggres
sion cannot make use of constructive energies; their zest for life is
crippled. The therapeutic task with such patients is to help them find
their destructive urge. As this clinical vignette demonstrates, patients
find meaning in constructive desires and making contributions only
after they fully experience their destructive impulses.
If the child is unable to "repair" the fantasized injury to the loved
object, there is no opportunity to resolve guilt over aggressive feelings
(Winnicott, 1963b). Like the character-inhibited patient, the depressive
is anxious over aggression and unable to use it constructively. The dif
ference is that the depressed patient, crippled by guilt, directs aggres
sion against the self. The typical depressive symptoms of self-hatred
and self-accusations are understood by Winnicott, as by Klein, as
manifestations of guilt over the intention to injure the loved object. In
its extreme form this dynamic results in "implosion," with rage being
directed with uncontrolled fury at the self in order to protect the loved
object (Winnicott, 1950). These patients run the greatest risk of suicide.
"Introversion" is also a regressive response to the failure of the
environment to contain aggression in the phase of relative depen
dence (Winnicott, 1950). This defensive maneuver also involves a type
of split, but in this case the good is kept within and all bad feelings are
projected onto external objects. A persistent pattern of such projection
defends against aggressiveness and provides relief from the guilt but
fills the world with persecutors, with the result that withdrawal is
then used as a defense against the world of hostile objects. The child
may attempt a recovery from introversion by attacking the persecutors.
D. W. Winnicott 179
In Winnicott's view, this "acting out" is a healthy sign, but if it is misun

derstood by the environment as a negative behavior, it will easily dis
appear, with the child lapsing back to introversion. In the extreme
form of this pattern, the true self is hidden away, and a false self is cre
ated to hide it. In this way, conflicts over aggression can lead to true
self-false self pathology.
As we have seen, Winnicott (1950) believed that aggressiveness
plays so crucial a role in the development of the self that if it is dis
avowed the growing child will not feel real. To gain a feeling of real
ity, outlets are sought for the obstacles that provide an opportunity for
an aggressive reaction. This is Winnicott's understanding of chroni
cally aggressive patients. Since the feeling of reality only lasts as long
as the expression of aggressiveness, such patients must continuously
do battle; they seek opposition over and over again, even over seem
ingly inconsequential events, in a desperate struggle to feel real by
aggressively opposing the environment. To an observer this type of
patient may appear to be hostile or to have difficulty controlling
aggressive impulses, but in Winnicott's view, the patient is attempting
to achieve a sense of reality. The therapeutic task is to understand this
meaning of the seemingly hostile behavior.
The traditional ego-psychological view of excessively aggressive
patients is that they lack drive control regulation owing to defective
ego structures. (In chapter 5 we will see that Kernberg put forward
such a view.) In opposition to this approach, Winnicott saw a self
defining meaning in rageful behavior. The clinical implication of Win
nicott's formulation that extremely aggressive patients are attempting
to achieve a feeling of reality and a sense of self is that such patients
need to be able to express their rage in the treatment setting. Expres
sion of aggression must be permitted for a variety of reasons: First, the
therapeutic environment must not repeat the mistake of the early care
givers by reacting negatively to aggressiveness; if therapists react by
blocking aggressive outbursts, they risk colluding in the reenactment
of the original trauma. Second, patients gain such sense of themselves
as they have by the expression of aggressiveness; a negative reaction
from the therapist robs such patients of their fragile sense of self and
leads to still more desperate efforts to achieve selfhood. Finally, and
most importantly, the hoped-for outcome of such expression is the
fusion of aggressiveness with the erotic drive, from which it was origi
nally de-fused as a reaction to trauma. In short, Winnicott felt that
since excessive hostility is due to the failure of the early environment
to contain the aggressive impulse, thus splitting it from the rest of
the personality, the task of the therapeutic setting is to "contain"
180 Chapter 4
aggressiveness in order to reintegrate aggression into the personality.

Containing the expression of aggressiveness is an example of the ther
apist's facilitating the resumption of the blocked maturational process
by acting as a "good enough" mother.
Patients who need objects for their aggression may desperately
attempt to "re-fuse" the erotic impulse with the aggressive drive
(Winnicott, 1950). In such patients the erotic impulse, instead of being
imbued with aggression, subserves aggression. In Winnicott's view,
both sadism and masochism may be understood in this way. The ori
gin of sadism lies in the need of the individual to feel real through
ruthless, destructive attacks on others that are exciting by virtue of
their fusion with the erotic impulse. To achieve this feeling, the object
must not be simply a "bad object" suitable for attack but the recipient
of erotic feelings. If the ruthless attacks are turned toward the self,
masochism results.
How far was Winnicott willing to go in allowing the expression of
the aggressive impulse? Little (1985) reported that Winnicott only
rarely attempted control over her many aggressive outbursts. He for
bade them verbally only when he felt that she was in danger—and he
was not always successful. He held her hands—to contain her anxiety,
not to control her aggressive outbursts. In one session, Little smashed
and trampled a vase; Winnicott's response was to leave the room and
return toward the end of the session. At the next session he had a new
vase. The incident was neither interpreted nor discussed at any point;
the only comment Winnicott made to her was that she had broken
something valuable to him. He took quick preventive action only to
protect Little. In one session, she intended to rush out of the office and
drive away recklessly. Winnicott took her car keys until the end of the
session. These incidents demonstrate that Winnicott allowed Little a
full range of aggressive expression within the limit of assuring her
safety. The protective measures he took strengthened the safety of the
holding environment until her destructive urge could be integrated
into her personality.
Winnicott's approach of allowing the expression of the aggressive
impulse holds for all pathological reactions to aggressiveness. His
treatment goal for splitting, reaction formation, and depression was to
bring the aggressive impulse back into "fusion" with the erotic desire
from which it was defused by environmental failure. The healthy per
sonality has most of the aggressive impulse fused with the erotic
drive, and the consequent ability to contain ambivalent feelings to
whole objects and to enjoy healthy love relationships (Winnicott,
1950). Even in health there is always an aggressive component "left
D. W. Winnicott 181
over" seeking obstacles for expression, but such aggressive expression

is experienced joyfully as "appropriate opposition" is sought. In order
for the patient to arrive at this level of development, the therapist
must find the split-off or hidden aggression and allow its fullest possi
ble expression in the safe environment of the treatment setting. Only
when the aggression is allowed such full expression can it be fused
with the erotic drive and integrated into the personality.
Aggression expressed outside the treatment setting was seen by
Winnicott as a reaction to the environmental failure of the therapy,
that is, as evidence that the patient found the setting unsafe for
aggression. He deemphasized the role of interpretation here, as he
did in the treatment of all preoedipal disorders. Since the treatment
objective is to correct an early environmental failure, therapeutic
action occurs by virtue of the provision of a safe environment for
the expression of the aggressive impulse and its fusion with the
erotic drive.
The phase of relative dependence marks the beginning of maternal
expectations and thus gives rise to yet another potential source of
pathology: the mother demands that the child be gratifying to her, the
child must learn to adjust what it feels in order to please her. In this
way, the mother molds such a child to please her at the expense of the
child's own feelings. Once again, the "kernel" has been sacrificed to
the "shell," and false-self pathology results. The child's aggression is a
major feature of this syndrome: if the child's aggressiveness is threat
ening to the mother, the child must defend against its expression in
order to allay her anxiety. If the mother is threatened by strong affect
in general, the child must bury all its feelings, which is to say, its
nascent personality. The result is a feeling of nonreality, leaving life
empty. Such a patient appears to be treatable by traditional analysis,
but treatment only begins when the true self is uncovered.
SUMMARY
Winnicott's theory of psychopathology and treatment is a direct out

growth of his concept of development as a series of phases of depen
dence that must be overcome to arrive at a healthy, independent adult
life. He saw all preoedipal forms of psychopathology as environmen
tally induced arrests in the movement from absolute dependence
toward independence, that is, from omnipotence to reality. Treatment
has to do with locating the developmental arrest and providing the
appropriate environmental response. For patients with preoedipal
psychopathology and for neurotic patients in moments of regression,
182 Chapter 4
psychotherapeutic action involves the provision of a therapeutic

atmosphere in which patients have the developmental experiences
they were denied in the preoedipal phase. The analytic process is
used, but it is modified according to the developmental issue alive in
the treatment at any given moment.
Winnicott proposed a hierarchy of values different from those in
the traditional concept of psychoanalytic therapy. His fundamental
premise is that psychoanalytic therapy is like the mother's role in
development: its essence is adaptation to need. Just as the mother
must understand the child's developmental level to provide for its
needs, so too must the psychoanalytic therapist gear interventions to
the patient's level of dependence and the associated developmental
blocks. Interpretations are used insofar as they fit the developmental
need, but adaptation is often the more appropriate modality. For
example, if the patient is unable to integrate aggression with the erotic
drive because of an early arrest in the experience of aggressiveness,
the therapist's role is to help bring aggression into the patient's
affection for the him or her so that the two impulses are "re-fused."
According to Winnicott, interpretations are most profitable when
understanding meets the developmental need, but even neurotic
patients frequently require other types of intervention. The tradi
tional model sees interpretation as the only mode of therapeutic
action; thus any need the patient appears to evince must be inter
preted (Brenner, 1979; Gill, 1981). From this viewpoint, if the aggres
sive and erotic impulses are unintegrated, walled-off aggression
would be interpreted, making the patient more aware of his or her
aggressiveness within the transference. The developmental level of
the conflict would also be interpreted to the patient, but any other
response would violate the theoretical framework. By challenging
the primacy of interpretation, Winnicott proposed a new, more
developmentally grounded, model for psychoanalytic therapy.
According to this alternative model, psychoanalytic therapy is the
application of a developmental understanding of psychopathology,
based on environmental failures encountered in phases of depen
dence, by means of the therapist's active adaptation to the patient's
childhood needs as they become manifest in the therapeutic setting.
Winnicott's thought bears a clear similarity to Guntrip's view of the
analytic process as offering a new beginning if the patient could reach
the most regressed portion of the psyche. The difference is that Winni
cott's more specific concept of development enabled him to gear treat
ment to the specific needs of individual patients, and thereby avoid the
homogenization of pathology that we observed in Guntrip's work.
D. W. Winnicott 183
As we will see in chapter 6, Winnicott's thought anticipated certain

features of Kohut's self psychology. Both theorists emphasized the
relationship between self and object as the key to development, both
recognized the importance of unrealistic expectations in early devel
opment, and both saw the development of the reality sense as a func
tion of gradual environmental failure. In the clinical arena, Winnicott
and Kohut were both sensitive to the importance of allowing patients
illusions of self-aggrandizem ent until they were ready for the
"descent" into reality. However, Kohut's developmental scheme, for
mulations of pathology, and treatment approach have certain essential
differences from Winnicott's, as will be discussed in chapter 6.
Winnicott's view of psychoanalytic treatment is not to be confused
with Franz Alexander's "corrective emotional experience." Alexander
(1950) artificially constructed situations to induce certain experiences
he thought would be therapeutic for the patient. Winnicott was
opposed to artificiality in the treatment process, believing that action
is therapeutic precisely because it adapts to a spontaneously
expressed need of the patient and that the art of therapy lies in the
ability to assess which developmental need is being manifested and
determining how to meet it. For the "false self" patient, artificial con
trivances are even more dangerous, since they fit into and buttress
existing defenses.
The result of therapeutic adaptation to spontaneously expressed
needs is an analytic space in which the patient creates a new self. It
is insufficient, in Winnicott's view, to overcome the past. Indeed,
it is probably not possible to overcome the past totally until new
experience is integrated into the psyche.
Clearly, in many clinical situations Winnicott would assess and
intervene differently from the manner prescribed by the traditional
psychoanalytic approach. Perhaps more than any other psychoana
lytic theorist, he based his interventions on the mother-infant rela
tionship. In this sense, he is the most purely object relational of all
analytic theorists.
CRITIQUE
Winnicott's theory is the most thorough, detailed, and consistent

version of the developmental arrest form of object relations theories.
The conceptualization of development in terms of phases of depen
dence has considerable observational and empirical support (e.g.,
Spitz, 1965; Mahler, 1975; Stern, 1980). Also impressive are the results
Winnicott frequently reported in his discussion of clinical cases. He
Chapter 4
made many theoretical contributions, but perhaps his most lasting

contribution is his belief that meaning can be sought in even the most
bizarre clinical material.
There are, however, certain conceptual ambiguities in Winnicott's
thought. First, the role of interpretation is not clear. If psychoanalysis
is adaptation to need, it is not clear why the process should ever con
sist of interpretation. Winnicott's answer is that interpretation is an
adaptation for some patients at certain times in the treatment. The dif
ficulty lies in understanding how interpretation could ever be an
adaptation to a developmental need. The child, after all, does not
require interpretation. The logic of Winnicott's view of the analytic
process as adaptation to arrested developmental needs would seem to
obviate interpretation. Since Winnicott clearly did not draw this con
clusion, the conceptual integration of interpretation in his treatment
model is an enduring problem.
Second, it is not clear exactly what adaptation to a developmental
need means. One cannot reverse time and begin the developmental
process anew. Regression is not a return to an earlier fixation point,
but the simulation of such a return (Bettelheim, 1972). It seems some
what mysterious that periods of regression in the consulting office,
often brief, could constitute the meeting of a developmental need.
Winnicott seemed to feel that the patient could "re-live" episodes and
even patterns in the analytic setting. But one may question whether
such re-living can actually occur and, even if it can, how exactly one
would know if and when it was occurring.
Beyond the problem of conceptualizing the re-living of earlier
episodes in general is the validity of certain inferences Winnicott
tended to make regarding the content of the re-living. This tendency
to make inferential leaps is most striking in Winnicott's not infrequent
interpretation that patients were re-living the birth process. For exam
ple, when Little was seized with terror during a therapy hour, Winni
cott told her that she was re-living her birth. Such an inference is, to
say the least, highly speculative, and one wonders on what basis it is
justified. Further, such an interpretation requires the assumption that
the patient remembers the birth experience, if not verbally, then
somatically. This assumption, too, is highly questionable.
In short, there are conceptual difficulties with Winnicott's theory.
Nonetheless, one cannot fail to be impressed with his clinical results,
at least for some patients. To achieve these results, he stretched the
concept of psychoanalysis considerably—his critics would say he dis
torted it beyond recognition. In fact, it is debatable whether his treat
ment approach ought to be called psychoanalysis. But more important
D. W. Winnicott 185
than the label is the effectiveness of his treatment strategy. Winnicott

found a way to reach some very disturbed patients, and this fact can
not be dismissed lightly. He exercised considerable influence on
a number of analysts who attempted to apply some of his concepts
in their own work. So whether or not Winnicott's work deserves
to be labeled psychoanalysis, it has given many analysts useful techni
cal principles and the courage to employ them in the treatment of
severely disturbed patients. We conclude this chapter with a brief
consideration of the work of some of Winnicott's closest followers.
THE "WINNICOTTIANS"
W innicott's followers have tended to emphasize three primary

aspects of his work: (1) the application of psychoanalytic therapy to
severe emotional disorder; (2) the use of the psychoanalytic setting
in this type of treatment; and (3) the role of the therapist and his
countertransference in the analytic process, especially in working
with severe pathology.
Khan (1963) believed that Winnicott contributed greatly to the
understanding of psychopathology with his concept of environ
mental impingement. In Khan's view, Winnicott added a critical
dimension to pathogenesis that is captured in the concept of the
"cumulative trauma." The primary maternal function is to operate as
a protective shield to keep impingements from interfering with ego
development. In agreement with Winnicott, Khan believed that the
infant needs a period of omnipotence before it can adapt to reality.
The protective shield allows the child to have this illusion, but if
there is a breach in it, trauma will accumulate silently as the child is
forced into awareness of the environment before the ego is ready to
manage reality awareness.
Children subjected to such cumulative trauma will be forced to
adapt to the environment by the formation of a false self, and the
development of psychological capacity will be arrested. The result is
an inability to use symbolism (Khan, 1962). Such children utilize
action to discharge tension states because they cannot use psychologi
cal means, such as language. The arrested psyche with little capacity
for symbolism and covered by a false-self organization is Khan's
description of the borderline personality disorder.
In Khan's view, the proper treatment approach is to recognize the
acting out of such patients as their only means of communication and
to tolerate it. The therapist acts as the protective shield so that the
patient can eventually tolerate the inner panic defended against in the
186 Chapter 4
acting out (Khan, 1964). The analyst operates as an auxiliary ego,

"holding" the situation until the patient can internalize the analyst as
a new object. The relationship with the analyst is not a transference in
the sense that the analyst is an object for the patient but the result of
the process is that the analyst becomes an object. Such patients, lacking
the ability to use symbols, communicate by what they make the ana
lyst feel, often conveyed in silence (Khan, 1963). Consequently, the
primary analytic instrument is the countertransference, and the ana
lyst's task is to pay close attention to his reactions to the patient
(Khan, 1960). The patient conveys an early primitive object relation
ship, and the feelings of the analyst may correspond either to the
patient's early experience or to the experience of the object. This
emphasis on the countertransference for understanding pathology
and reversal of roles in the transference-countertransference enact
ment is the link between Khan's Winnicottian approach and the
Kleinian school.
In Khan's view, Winnicott's concept of illusion, the transitional
space between reality and fantasy, describes accurately the psychoan
alytic space Freud created with his discovery of "cure by language"
(Khan, 1971). Psychoanalysis creates an area of illusion so that sym
bolic discourse can take place. The borderline patient is unable to use
this space and infuses it with action. The goal of treatment for such a
patient is to gain the ability to utilize the analytic space through sym
bolic discourse. Since patients who possess symbolic capacity use the
analytic space to create a new sense of self, the analyst also functions
for these higher level patients as a vehicle for new self experience. In
Khan's view, Freud's genius was in creating a setting in which new
forms of self experience could be created, and Winnicott's brilliance
lay in his ability to see Freud's work in this way.
M argaret L ittle, whose analysis with W innicott has been
described, also applied Winnicott's insights to severely disturbed
patients. Little (1981) drew a critical distinction between transference
neurosis and the delusional transference. Like Khan, she believed
that the former is a relationship based on symbolism whereas the lat
ter is an undifferentiated state in which words have little meaning.
In Little's view, borderline and psychotic patients, who form delu
sional transferences, lack integration of psyche and soma. Such
patients exist on the action level; they must act because their somatic
existence is disconnected from the linguistic, or psychological, level.
Consequently, they cannot use interpretations but require the pre
sentation of reality, and the delusional transference must be broken
up before interpretations are useful.
D. W. Winnicott 187
Little advocated two primary deviations from classical technique to

dissolve the delusional transference so that interpretations could be
employed. Like Khan, she believed that the use of the countertransfer
ence was critical to the treatment of severely disturbed patients who
communicate experience by evoking feelings in the analyst. Little felt
that in the countertransference lies the key to understanding. The aim
of treatment in severe character pathology is not to understand the
symbolic meaning of verbal communication, as it is with neurotic
patients, but to understand the meaning of the analyst's reactions
to the patient. In Little's view, if these patients are to let go of their
transference delusions, the analyst must, at the appropriate point in
treatment, let them know his or her feelings and reactions. This com
munication constitutes Little's first deviation from classical technique
and its purpose is to help the patient differentiate reality from fantasy.
Often the patient's perceptions of the analyst are accurate, and the
patient needs to know this to appreciate the parents' real irrationality.
Also, gaining knowledge of how the analyst feels can shock the
patient into reality.
Little's second deviation from classical technique is her occasional
use of physical intervention to help the patient dissolve the delusional
transference. If Little felt there was no other way to "present reality"
to the patient, she would make body contact, usually by reaching over
and touching the patient. The purpose of this maneuver was to help
the patient differentiate self from other. For example, Little put her
hand on the ankle of a female patient who felt entrapped and let
her push it away; the patient felt free, and this event marked the
beginning of her sense of differentiation.
The breakup of the delusional transference, in Little's view, is simi
lar to awakening from a dream. The patient must be confronted with
reality, and the analyst's "personhood" and presence is the reality
with which the patient must be confronted. It is up to each therapist to
find the way to use his or her own reality to break up the delusional
transference. Once it is dissolved, the patient advances to a symbolic
level and can make use of interpretations.
For Little's technique to be effective, the analyst must respond from
within, as a real person. Little not only saw the countertransference as
a crucial component of the analytic process but also broadened the
concept to embrace the analyst's total response to the patient.
Although she agreed with Winnicott that there is an "objective" coun
tertransference consisting of the analyst's reactions evoked by the
patient, she also believed that the countertransference could not be
limited to such feelings. The treatment of character-disordered
188 Chapter 4
patients, the group in which Little included most of her patients,

required, in her view, a commitment from the analyst to go as far
"into the patient's illness" as possible. Deprived patients must be
given to before they can give, and what analysts can give is limited
by who they are as a people. The patient's deepest needs cannot
really be met; analysts can only offer themselves with their realistic
strengths and limitations. However, these very limitations are not
drawbacks but assets, since they make the reality of the analyst
known to the patient.
In Little's case of Frieda, seven years of transference interpretations
had no meaning until the patient's friend died and, Little expressed
her feelings of sorrow, distress and pain for the patient in her grief.
This open acknowledgement of feelings was the turning point of the
analysis, since the patient felt for the first time that the analyst was "a
real person" and not "counterfeit" like her parents. Thereafter, trans
ference interpretations began to have meaning to her. It was Little's
contention that if she had not genuinely felt distress and pain on hear
ing of the friend's death, her expression of feelings would not have
been effective.
Andre Green is the other major contributor to psychoanalysis
whose views grow out of Winnicott's thought. Green (1978) pointed
out that psychoanalysis has refound subjectivity in object relations
theory. He noted that in Freud's metapsychology the object is linked
to the drive, not the subject. In object relations theories, however, the
complement of the object is not the ego but the self, making the sub
ject as an experiencing person the central focus of the analytic
process. In Green's (1978) view, Winnicott made several lasting con
tributions to psychoanalysis: (1) he made psychoanalytic theory
applicable to borderline personality disorders, which, Green felt are
the paradigmatic cases of our time; (2) he put the countertransfer
ence in the center, not the periphery, of the treatment process; (3) he
gave theoretical focus to the psychoanalytic setting; and (4) he made
the patient's experience, rather than theoretical schematizing, his
primary concern.
According to Green, Winnicott recognized that a crucial aspect of
development is the evolution of the child's sense of the object from the
"subjective object" to the "objective object" and that the role of the
mother is crucial in this process. Borderline and other severe character
pathology is characterized by arrest in the transitional phase between
the delusion of the "subjective object" and the recognition of objective
reality (Green, 1975). The implication for treatment of such patients is
the use of the "analytic space" for the creation of a new object. The
D. W. Winnicott 189
analyst is a participant in this endeavor, since the new object is created

out of the therapeutic interaction, just as early objects in normal devel
opment are created out of the mother-child relationship. In Green's
view, Winnicott made a seminal contribution with his reconceptual
ization of the role of the analyst from detached observer to a partici
pant in the process. The aim of treatment is to make the "analytic
object," which is created out of the therapeutic interaction, usable by
the patient in the analyst's absence. In Green's view, among all psy
choanalytical theoreticians, only Winnicott conceptualized the analytic
setting as potential space for the creation of a new object and this con
cept describes the analytic process and its goals for today's patient
more accurately than does Freud's notion of the blank screen on
which templates from the past are projected.
The pathology of the borderline personality disorder is conceptual
ized by Green (1977,1978) as an absence. While neurosis may be usefully
conceptualized in terms of conflictual mental representations, the key
feature of the borderline personality is the very lack of object representa
tion, resulting in a blankness, a lack of self development. According to
Green, treatment of such a patient cannot be focused upon "dis-covery"
because no object is to be found. In Green's view, it was Winnicott's
unique contribution to understand that the conceptualization of transfer
ence as repetition of early object relations is not applicable to character
pathology. Rather, the essence of treatment must be the transformation
of absence into potentiality by the creation of the "analytic object." Green
believes the future of psychoanalysis rests in this conceptualization
because it applies to the most typical cases now seen in treatment.
Green (1975) agreed with Little that the split between psyche
and soma makes borderline patients unable to use symbols. Such
borderline patients communicate either by action, acting out, or
somatization, and the therapeutic challenge is to translate this type
of language into verbalization. In Green's (1977) view, the primary
tool in this process is the countertransference. His view, in agree
ment with Little's, is that therapists must use the feelings induced
in them by borderline patients as the key to understanding. The
other primary technical instrument used by Green is the analytic
setting, which, if properly employed, can become a metaphor for
maternal care. Since the provision of a "good enough" environment
is believed to facilitate the use of symbols, Green, like Little, sup
ported variations in technique for the purpose of fostering the use
of symbolization. In the analyst's efforts to use himself or herself to
facilitate symbolization, analyst and patient alike will come upon
the analyst's own limitations, but in Green's view, these limitations
190 Chapter 4
are a crucial part of the treatment, as they aid patients in their journey
toward the recognition of reality.
Green's views have much in common with those of Little and, to
a lesser extent, of Khan. All three Winnicottians see Winnicott's last
ing contribution as his reconceptualization of the analytic process as
the unblocking of early developmental arrest by the creation of a
new object and its corollary, a new self. Analysis becomes an experi
ence with a new person that allows for the transcendence of old
perceptions, patterns of relating, and defenses.
Thus, W innicott and his follow ers shift Freud's model of a
detached scientist observing and providing insight to a model of two
selves in interaction. Both participants in the exercise are affected, and
both are undoubtedly changed. The outcome is successful if the
patient becomes a new person, experiencing himself or herself and
others in new and richer ways.
CHAPTER 5
The Work of Otto Kernberg
O tto K er n ber g ' s c o n c e p t u a l f r a m e w o r k is c h a r a c t e r iz e d b y a b l e n d

of object relations theory and ego psychology. Kernberg (1972) wrote
a thorough critique of the Kleinian school, emphasizing its violation of
ego psychological principles of theory and technique and suggesting
that Klein's object relations approach to development and psy
chopathology be integrated with the findings of ego psychology.
Kernberg's work may be looked upon as an effort to achieve the goal
he outlined in that critique. As we shall see, Kernberg utilizes Klein's
theoretical tools in his conceptualization of severe personality disor
ders, but he makes major modifications in Klein's theory of develop
ment and does not adhere to K lein's principles of technique,
preferring to focus interpretation on specific types and levels of
defense as described in ego psychology.
Kernberg offered several major criticisms of Klein's views, three of
which were discussed in chapter 3; namely, Klein's adultomorphism,
her pathologizing of development, and her assaultive interpretive
approach. In addition, Kernberg attacked Klein for failing to appreci
ate the importance of diagnosis and the need to adapt treatment to the
type and severity of the disorder and for her disregard of environ
mental factors, neglect of psychological structure, conceptualism of
defensive organization and developmental process, and ambiguity in
terminology. Despite all the weaknesses he finds in Kleinian theory,
Kernberg accepts the importance of early object relationships, aggres
sion, envy, splitting, and the projective-introjective processes in both
early development and the pathogenesis of character pathology.
Kernberg also adopts Klein's view of the defensive constellations against
aggression and envy and her notion of early superego formation.
The other major influence on Kernberg's object relations psychol
ogy was Edith Jacobson. From Jacobson, Kernberg (1980) takes the
concept of ego development as a product of object relationships. As
discussed in chapter 1, in Jacobson's (1964) view, life begins with an
initial undifferentiated phase, out of which object-representations of
191
192
the self and objects gradually become differentiated in the second

phase, although they are still split between "all good" and "all bad."
The third phase involves the integration of the self- and object images
to form psychic structure and affects become intrapsychic regulators
by means of their investment in representations of the self and objects.
Similarly, Jacobson believed that the superego was formed from the
integration of the initially split bad-object images with the ego-ideal
and parental prohibitions. Kernberg builds on Jacobson's develop
mental integration of object relations theory and ego psychology to
construct his general theory of psychopathology.
Kernberg (1980) was also influenced by Mahler's theory that after
the child emerges from the symbiotic phase, it enters the separation-
individuation process, which ends in the oedipal phase (Mahler et al.,
1975). Kernberg's view concurs with Mahler's formulation that pre
oedipal development includes a gradual process of separating from
the maternal figure and individuation of the sense of self and that
these issues influence oedipal development. However, Kernberg's
view of severe character pathology tends to emphasize Klein's con
cept of splitting more than Mahler's subphases of the separation-
individuation process.
METAPSYCHOLOGY
Although Kernberg, like Klein, adheres to the dual-drive theory, the

primary motivational system for Kernberg consists of inborn affect
dispositions. All gratifying and frustrating experiences stimulate
affect, which is initially pleasure and unpleasure. Early affect states
are organized in this very rudimentary way, but as perceptual and
cognitive elaboration increases, affective states become increasingly
more complex and their discharge function becomes less important.
Since the child's experiences with the environment trigger its affective
states, the affect is always embedded within a relationship between
self- and object images, however rudimentary they may be. These
object relations units are stored as "affective memory." The experien
tial store of psychological experience upon which the psyche is con
structed consists, therefore, of object relations units, each of which
consists of a self- and object image and a connecting affective link.
Kernberg (1976) makes an effort to integrate psychoanalytic drive
theory with the modern concept of instinct, pointing out that the trend
in both ethology and neuropsychology is away from the view of
instincts as given responses to specific environmental circumstance
and toward a concept of instincts as "organizations which, through
193
learning, integrate various inborn patterns ('building blocks') into

flexible, overall plans"(p. 86). He notes that Freud differentiated the
biological Instinkte from Trieb, or drive, which lies on the borderline
between the biological and the psychological. Instincts are discrete,
self-preservative behavioral responses called upon when necessary.
Drives are cyclical, initially physiological stimulations to action that
motivate behavior for their gratification and therefore give rise to
wishes, their psychological expression. Freud's Trieb is, for Kernberg,
an organized system built on the initially discrete "instinctive compo
nents," or "building blocks," that are "released" by the mother-child
interaction. Kernberg's concept that drive develops in the context of
the mother-child relationship is similar to Loewald's (1971) view.
Kernberg argues that the innate, instinctive components gradually
develop into the drive organization as pleasure becomes a component
of libidinal object relations units and "unpleasure" evolves into
aggressive object relations units. In accordance with this shift, affects
change their function as they and the developing drive system become
more complex; at first organizers of the instinctive components, affects
become "signals" of the drive organization.
One can see from these conceptualizations that for Kernberg drives
are not inborn motivational units, as they are for Klein. Affective dis
positions and instinctive, behavioral "building blocks" are inborn, but
the drive organization is conceptualized as an outcome of develop
mental experience that begins as undifferentiated object relations
units and develops gradually into more complex object relationships.
In its psychological form, a drive is represented by a wish for an
object. Given this, one might conclude that object relationships replace
both drives and affects as the primary m otivational system in
Kernberg's theory. This is just the conclusion drawn by Fairbairn and
Guntrip (see chapter 2). Kernberg, however, chooses not to view
object seeking as primary.
Kernberg (1976) offers three reasons for his opposing Fairbairn's
and Guntrip's replacement of drive theory by an object-seeking
model. First, the Fairbairn-Guntrip view ignores the importance of the
aggressive drive, which seeks not to attach to an object but to elimi
nate it. Second, the organization of the psyche into "good" and "bad"
object images is more important than the fact that the contradictory
affects are directed to the same object. Finally, making object striving
primary ignores qualitative shifts in the drives during crucial develop
mental advances, such as the transformation of libido into the genital
organization in the oedipal phase; that is, the striving for the breast is
not the same as genital wishes, a difference that Kernberg believes is
194 Chapter 5
abolished by the primacy that Fairbairn and Guntrip place on object

seeking. For these reasons, Kernberg's view differs from those of
Fairbairn and Guntrip and maintains the primacy of drives in devel
opment; Kernberg views affect, not the object relationship, as the pri
mary motivator of the psyche. It should be emphasized, however, that
since affects are always experienced within an object relationship, in
Kernberg's formulation, the two are difficult to differentiate.
Like Klein, Kernberg (1976) assumes that positive early experi
ence leads to libidinal object relationships and negative experience
results in aggressive object relationships. Unlike Klein, but in accor
dance with Jacobson, he believes that the development of these
object relationships results in the structure of the ego, which deter
mines the health or pathology of the personality. As opposed to
classical ego psychology, which views ego structure as a product of
drives and their vicissitudes, Kernberg's theory emphasizes that ego
structure results from object relations units, which also determine
the drive organization. Development, for Kernberg, consists of a
series of internalizations of object relations units and the defenses
against them. The dynamic unconscious, therefore, consists of object
relations units defended against, and hence made unconscious, by
either primitive early defenses or more advanced later defenses.
Whereas for Klein and her followers psychological structure is tan
tamount to object relations fantasies, for Kernberg, as for main
stream theorists, there is a clear distinction between fantasy and
structure: the object relationship leads to psychic structure but is
not equivalent to it.
The internalization of object relationships is of three types in
Kernberg's (1976) scheme. The most primitive form of internalization
is introjection. Whole interactions with the environment are taken into
the psyche and reproduced with the corresponding self- and object
image and affective coloring. In the next phase the child begins to
identify with the parent as the developing ego takes in the role aspect
of social interaction. In Kernberg's view, the child's imitation of the
mother and the performance of roles constitute the first identifications
and the intermediate level of internalization. The highest level of
internalization in this scheme falls under the rubric of ego identity,
which Kernberg (1976) defines as "the overall organization of iden
tifications and introjections under the guiding principle of the
synthetic function of the ego" (p. 32). These three modes of psycholog
ical growth and organization occur in developmental sequence and
we shall now turn to this development view of the internalization
of object relations.
Otto Kernberg 195
DEVELOPMENT
Kernberg (1976) divides development into five stages. In the first

month (Stage 1), the normal undifferentiated self-object representa
tion is built. From the second month until about the sixth to eighth
month (Stage 2), self- and object images become differentiated. Also,
the consolidation of the "good" self- and object representations occurs
in response to pleasurable, positive m other-child interactions.
Simultaneously, all negative experiences become organized into "bad"
self- and object representations. These two separate object relations
units constitute the first introjections. The good self- and object repre
sentations become invested with libido whereas the bad self- and
object representations are invested with aggression. Initially, these
two kinds of object relations units are separate because of the physio
logical inability of the primitive ego structure to integrate them.
Beginning at about the third to fourth month the ego actively splits
the two kinds of self- and object images apart in order to protect the
"good" self- and object images from the destructiveness of the bad
images. At this point splitting becomes the primary defense of the
developing ego. This stage encompasses Mahler's symbiotic phase, in
which self and object are not yet distinguished, and her differentiation
subphase of the separation-individuation process. Kernberg collapses
the two phases in his developmental schema because he believes that
in the differentiation subphase refusion of self- and object images
takes place frequently.
The third stage begins with the differentiation of self- and object
images within the good and bad object relations units and ends with
the integration of both good and bad self-images and good and bad
object images. It begins at about six to eight months and lasts until 18
to 36 months, when the integration of self- and object images results
in object Constance. At the beginning of this phase re-fusion of self-
and object images still can happen under stress, but during this phase
ego boundaries are sufficiently well established that they are main
tained even under stressful conditions. In this schema, just before or
during the third year the child begins to integrate good and bad
images in both his self- and object representations. By the end of this
phase, good and bad self-representations are consolidated into an
integrated self-concept, and good and bad object images form total
object representations.
Stage 4 extends throughout the oedipal period and is characterized
by the solidification of libidinally and aggressively invested self-
images into "a definite self system" and good and bad object images
196 Chapter 5
into "total object representations." Ego, id, and superego become dif
ferentiated as defined psychological structures. While one can see
Kleinian influence in Kernberg's emphasis on early splitting and the
development of self and object integration, in contrast to Klein,
Kernberg believes that the integration of self- and object representa
tions in this phase means that repression replaces splitting as the pri
mary defensive organization, thus allowing for the division of the
psyche into ego, superego, and id. Whereas for the Kleinians repres
sion is pathological only if it is related to unresolved splitting, for
Kernberg the appearance of repression involves the establishment of a
new defensive organization that can be pathogenic in itself. Negative
introjects are no longer split off into a separate ego organization but
are repressed within a unitary psychological structure. It is only at this
point that psychological structure is formed and an ego organization
can be meaningfully spoken of. In this way, Kernberg integrates
Klein's object relations theory with the ego psychological emphasis on
structure. In Kernberg's formulation, ego identity is established in this
phase as a result of the integration of self- and object representations.
According to Kernberg, the integration of self- and object images
gives rise to ideal self-representation and ideal object representations.
Following Jacobson, Kernberg believes that these ideal representations
must be integrated with the earlier fantastical, sadistic superego fore
runners. Under optimal circumstances the consolidation of these
sadistic, attacking superego precursors with the new ideal-representa-
tions and with realistic parental prohibitions leads to the superego as
an independent psychic agency.
The final stage of development, Stage 5 in Kernberg's formula
tion, is the consolidation of ego and superego integration. As the
superego becomes consolidated, its sharp distinction from the
ego decreases and the superego is gradually integrated into the
personality. The integration of the superego reciprocally fosters
ego identity, which also continues to develop and solidify through
effective relations with others, which are in turn made possible by
the already achieved integration of the psyche. The modulation of
object images in the previous phase allows for satisfying relations
with external objects, which, in turn, help solidify both self- and
object images. If all goes well in this phase, relationships with the
world continue to solidify the integrated self- and object images
and with them, the integration of ego and superego.
In Kernberg's developmental scheme two key steps are crucial
to avoiding the development of severe psychopathology. First,
self-object differentiation must take place. The failure of this psychic
Otto Kernberg 197
reorganization corresponds to the psychotic potential of the develop

ing personality. When self-object boundaries become established, the
psyche is defined as separate from the environment, but the psycho
logical organization continues to be defined by splitting and its
related defenses. The second crucial developmental step is the shift
from splitting to self- and object-integration, which, for Kernberg, is
equivalent to the shift from a split ego to an integrated ego that orga
nizes defenses around repression. Failure to achieve this developmen
tal milestone results in ego weakness and severe character pathology.
Once repression and related defenses replace splitting as the primary
organizer of the psyche, the ego is integrated and neurosis is the worst
possible outcome.
PSYCHOPATHOLOGY
These stages of development, for Kernberg (1976), demarcate the

developmental issues of the major categories of psychopathology.
Fixation in the first stage results in the failure to develop the undiffer
entiated self-object image and inability to form a symbiotic bond with
the maternal figure. The resulting pathology is autistic psychosis.
Personality arrest in the second stage is reflected in the lack of differ
entiation of self-object boundaries; the child never emerges from, or
regresses to, the symbiotic phase, when self- and object-representa-
tions were merged. Here Kernberg places adult schizophrenia, psy
chotic depression, and childhood symbiotic psychosis. Although
Kernberg mentions these disorders in the context of his overall theory
of development and pathology, he does not discuss pathological fixa
tions in Stages 1 or 2 and offers no contribution to their understanding
and treatment. Kernberg makes the disorders originating in Stage 3
the focus of his theory of psychopathology.
Kernberg proposes a classification of character pathology based on
the developmental shift from splitting to repression. "Low level" char
acter disorders are those organized around splitting and related
defenses. Character pathology at this level has its fixation point in Stage
3, as positive and negative self- and object-representations have not
been integrated. There is so little superego integration that superego
"nuclei" tend to be projected, resulting in paranoia as a major character
trait in these patients. According to Kernberg, since their character traits
are "instinctually infiltrated," such patients tend to be impulsive.
Because splitting is the primary organizer of the psyche, there is little
ego integration or object constancy. Kernberg includes in such low level
character pathology the borderline personality organization, many
Chapter 5
narcissistic personalities, sexual deviancy, hypomanic disorders;

most infantile personalities; antisocial, impulse-ridden, "as if," and inade
quate personalities; and prepsychotic character disorders. In all these
character disorders, which share the use of splitting and related defenses,
pathology is defined by the failure to develop the integrated object
relationships out of which an integrated ego and superego emerge.
In the intermediate and higher level character disorders, where the
integration of self- and object representations has been achieved and
object relationships are stable, pathology results from conflict between
ego and superego structures.
In the intermediate level of character pathology Kernberg places all
character disorders that primarily utilize repression, but that use some
lower level defenses such as splitting as well. At this level the super
ego is severe, but it is not well integrated; that is, the harsh sadistic
superego exists alongside the primitive ego ideal. Intermediate level
character patterns involve fewer inhibitions and more impulsivity
than higher level character disorders and some degree of splitting.
Kernberg includes in this category passive-aggressive, sadomasochis
tic, infantile, and many narcissistic personalities. In all these personal
ity disorders the defenses organized around repression are used, but
not exclusively. The clear separation of instinct and defense character
istic of pathology in the high level character disorders is lacking here
because the ego structure is neither as well organized nor as stable.
Kernberg believes psychoanalysis is the treatment of choice for these
disorders but acknowledges that the treatment will be lengthy and
difficult and may involve modifications in some cases.
Character pathology organized around repression and related
defenses, such as suppression, isolation of affect, and intellectualiza
tion, is high level pathology because the ego has an organization that
clearly separates defense and drive. Character defenses at this level
tend to be phobias, inhibitions, and reaction formation. Kernberg
includes in this category hysterical, obsessive-com pulsive, and
depressive-masochistic personality types. Here, character pathology
has advanced beyond Stage 3 in his schema to the structuralization of
the ego. The superego has become consolidated as a separate psychic
agency, but it is not integrated into the ego. Consequently, its severity
inhibits affects and personality development in general. The result is
an organized ego that can function but at the cost of diminished emo
tional gratification, at best, or severe depression and complete inabil
ity to experience satisfaction, at worst. Because conflicts take place
within an organized psyche, the treatment of choice for the highest
level of character pathology is psychoanalysis.
Otto Kernberg
It should be noted that narcissistic personalities are placed at differ

ent levels of Kernberg's schema, resulting in some confusion regard
ing his categorization of these cases. Pathological narcissism is placed
on a spectrum ranging from Stage 3 to 4. Kernberg categorizes most
narcissistic cases in Stage 4, with regression to Stage 3. He views these
cases as fluctuating between the two stages, since the majority of nar
cissistic personalities are not completely fixated in Stage 3 like patients
with borderline personality organization, but have not advanced to
complete self- and object integration.
It can be seen from this discussion that Kernberg's classification
of character pathology has direct treatment implications. The deci
sion to conduct psychoanalysis proper or expressive psychother
apy depends on the diagnosis of character pathology. Since these
treatments have distinct attributes, diagnosis, for Kernberg, is a
task of considerable importance. Consistent with his structural
understanding of personality and pathology, Kernberg focuses his
diagnostic approach on the organization of object relationships.
The patient's psychological structure is reflected in the degree of
identity integration, characteristic defenses, and capacity for reality
testing. The interviewer seeks to determine if the patient is domi
nated by introjection, has some degree of identification, or has
achieved ego identity. The diagnostician is especially attuned
to the determination of whether self-object differentiation has
occurred and, if so, whether the defenses are organized around
splitting or if a significant degree of ego and object integration has
been achieved. To accomplish these goals, the interviewer has three
diagnostic tasks: he must explore the patient's subjective world,
observe the patient-interviewer interaction, and observe his or her
own reactions.
The content of the interview is initially conducted along the lines
of the standard assessment interview, but when symptoms appear
during the interview, Kernberg (1984) recommends that the in
terviewer utilize the three therapeutic techniques: clarification,
confrontation, and interpretation. Clarification is the request for
explanation of what is unclear or contradictory at the conscious or
preconscious level. In confrontation, the interviewer directs the
patient's attention to information he or she has presented that is
contradictory and to the presence of conflictual functioning and
relates these observations to other areas of the patient's life and
functioning. Interpretation is the offering of possible unconscious
motivations for the perceived areas of conflict. In this process the
interviewer builds up a view of the patient's self- and object images
200 Chapter 5
that allows a placement of the patient into one of the three broad
categories of personality organization or, if the patient is character
disordered, into one of the subgroups of character pathology.
The value of Kernberg's comprehensive diagnostic system and con
ceptualization of development and psychopathology lies in the appli
cation of his object relations/ego psychology integration to severe
personality disorders, especially the borderline and narcissistic orga
nizations. Like Jacobson (1964) and Mahler (1975), Kernberg applies a
psychoanalytic approach to the "intermediary" area of developmental
experience between merger and the formation of psychic structure, an
area that he feels accounts for pathology between psychosis and neu
rosis. In this way, Kernberg contributes to what Stone (1954) referred
to as the "widening scope of psychoanalysis."
The Borderline Personality Organization
Borderline pathology, according to Kernberg (1975), is in essence the

failure to resolve the critical task of Stage 3 to integrate good and
bad self- and object images. Borderline patients, in this view, have
not been able to blend these two types of images into an integrated
self-concept or whole-object representations. Either because of
inborn excessive aggressiveness and inability to tolerate frustration
or because of severe trauma in this phase leading to excessive
aggressiveness, the good self- and object representations are continu
ally threatened by bad representations of both types resulting in
splitting the good and bad self- and object representations. The inte
gration of the ego, which leads to the tripartite division organized
around repression as the principal defense, does not develop, and
the result is general ego weakness, as the ego structure necessary to
manage tensions and conflicts between drives, superego strictures,
and the environment is missing. Rather than the structured ego that
is the normal outcome of Stage 3, one sees a fixation at the primitive
defensive organization of the splitting phase, including projection,
projective identification, and introjection.
This primitive defensive organization can result in fusion of self-
and object images in close relationships, leading to transient psychotic
episodes. Nonetheless, reality testing remains intact in other situa
tions, since self- and object images have become differentiated in the
previous phase. This defensive constellation organized around split
ting and the tendency to fuse in close relationships while reality
boundaries are maintained in other situations defines the borderline
personality organization in Kernberg's view. He believes that most
Otto Kernberg 201
cases falling within this category require psychoanalytic therapy with

parameters in order to control severe, frequently psychotic transfer
ence regressions. In his more recent writing Kernberg has expressed
more optimism regarding unmodified psychoanalytic treatment for at
least some borderline patients (Kernberg, 1984).
Kernberg (1975) enumerates the following as typical symptoms and
character patterns of the disorder (although no single symptom is
pathognomonic): chronic, diffuse anxiety; polysymptomatic neurosis;
polymorphous perverse sexual trends; paranoia; schizoid personality;
hypomanic and cyclothymic personality; infantile, "as if," antisocial,
and narcissistic personalities; and impulse neuroses (Kernberg, 1975).
Some combination of these symptoms is present in all borderline per
sonality organizations, resulting in the chaotic, confusing, disorga
nized clinical presentations so characteristic of these patients. The
presenting features suggest a diagnosis of borderline personality orga
nization, but in place of the DSM-III approach to diagnosis by the enu
meration of symptoms, which has no apparent relevance to treatment,
Kernberg (1984) proposes an assessment of the psychological structure
of the patient, which he believes has direct treatment implications.
The manifest symptoms, some of which will appear in every border
line personality, are, in his view, a product of an ego structure that is
not simply defective but is based on pathological internalized object
relations and the use of splitting as an active defense to keep apart
good and bad object representations. Thus, Kernberg uses this fixation
point to explain virtually all the symptoms and character features
typically found in the borderline personality organization.
Kernberg's structural analysis of the borderline personality organi
zation includes four structural indicators: "non-specific manifestations
of ego weakness," splitting, a shift toward primary-process thinking,
and pathology of internalized object relationships. Kernberg attributes
the nonspecific manifestations of ego weakness to the split between
libidinal and aggressive self- and object images that prevents the inte
gration of the ego structure necessary for the management of anxiety
and impulse. Here Kernberg invokes Hartmann's (1939) concept (dis
cussed in chapter 1), that libido neutralizes aggressiveness and inte
grates it into psychic structure. When the two drive organizations are
split, the ego cannot be structuralized and lacks the ability to manage
anxiety and impulse. Since the specific defense mechanisms of border
line pathology—splitting, primitive idealization, denial, devaluation,
and omnipotence—are organized around splitting object relation
ships, this second structural indicator is tantamount to the characteris
tic object relationships of the disorder. Kernberg attributes the "shift
202 Chapter 5
toward primary process thinking" primarily to the pathology of inter

nalized object relationships and the "reactivation" of splitting and
related defenses, which, as we have seen, is pathology of internalized
object relationships; thus, this structural indicator is a product of
regression in object relationships. The fourth structural indicator,
pathology of internalized object relationships, is tantamount to split
ting. Thus, Kernberg's four structural characteristics of the borderline
personality organization are all attributable to splitting good and bad
self- and object images and its consequences.
Splitting is illustrated by Kernberg's brief description of Miss A
(Kernberg et al., 1988). The patient viewed the therapist in two dis
tinct ways: as a strict parent and as a loving, tolerant person. When
the therapist was seen as the harsh object, Miss A experienced her
self as depreciated and hated herself and her body. When the ther
apist was a positive, caring object-representation, she felt caressed
and loved and felt free to express her wishes to exhibit her body.
Each object image was linked to its corresponding self-image by an
affective state, but the two self- and object image units were not
experienced by the patient as belonging to the same self or object.
This separation of the two disparate views of self and object is the
clinical manifestation of the use of splitting as a defense. Miss A's
feelings of fear of the object and her self-depreciation and hatred
were too extreme to be connected to the positive feelings she also
had about the therapist and herself; she feared that she would be
damaged, if not destroyed, by the contact.
The conclusion to be drawn from this account of the borderline per
sonality organization is that the chaos and confusion of the presenting
clinical picture is not simply an indication of an ego defect. Ego psy
chologists have tended to view the deficits seen in the borderline per
sonality solely as "weaknesses" in the structure of the psyche. While
Kernberg fully recognizes the defects in the ego of the borderline
patient, he considers them to be an outcome of the active use of split
ting to keep "good" and "bad" self- and object images apart. In other
words, what appears to be chaos betokens an active defensive process.
In this sense, Kernberg, like Winnicott, is a critic of the purely ego-
psychological approach to severe character pathology, as he believes
the description of ego defects does not provide a sufficient explanation
for the structural organization of the psyche.
In Kernberg's view other characteristic defense mechanisms of the
borderline personality are products of the split-ego organization.
Because the good object representation is unable to withstand contact
with the bad object representation, it must be free from all negative
Otto Kernberg
features and all aggressiveness, so that it easily becomes an unrealis

tic, all-powerful, all-good object image. Kernberg (1975) differentiates
this "primitive idealization" from later idealizations based on reaction
formation that are motivated by the need to defend against aggres
siveness. Primitive idealization involves no aggressiveness, con
sciously or unconsciously; it is a "primitive fantasy structure in which
there is no real regard for the ideal object, but a simple need for it as a
protection against a surrounding world of dangerous objects" (p. 30).
Another consequence of split self- and object images is that the
patient is left with excessively aggressive images of both types, unmit
igated by positive contact. The pain of these images leads to the need
to project them; consequently, paranoid trends are strong in the bor
derline patient. The immediate consequence of the projection is fear of
retaliation from the now dangerous "external" object that contains the
projected aggressiveness. Kernberg calls this an "early form" of pro
jection. Since the projection tends to be unsuccessful owing to the
weak ego boundaries, borderline patients utilize projective identifica
tion by identifying with the object onto whom their aggressiveness
has been projected, exacerbating the fear of their own projected
aggressiveness. To protect themselves, these patients attempt to con
trol the object in order to avoid attack, further weakening ego bound
aries. In close affective relationships patients alternate self- and object
images so quickly and intensely that reality testing becomes lost and
regression to psychosis may occur within the context of these relation
ships while reality boundaries are maintained in all other situations.
This is why loss of self-object boundaries is common in these patients'
close relationships. This type of projection is to be contrasted with its
forms, found in neurotic pathology, in which ego boundaries are
delineated so that there is no need to control the projected affects and
projective identification is unnecessary. As shall be seen later, the loss
of reality boundaries in borderline patients has critical implications for
the transference of these patients.
Borderline patients use denial to maintain splitting, since they must
be able to "mutually deny" two areas of consciousness. Denial may
manifest itself as a simple disregard of entire areas of subjective expe
rience or as actual denial that they exist. If pressed, patients may be
able to acknowledge that what they experience now contradicts what
they felt and said previously, but the denied sector of the personality
has no emotional relevance.
The final defenses Kernberg lists as characteristic of the borderline
personality organization are omnipotence and devaluation. Again,
Kernberg (1975) acknowledges that these defenses are also "intimately
204 Chapter 5
linked to splitting" (p. 33). The identification with the all good self-
image easily results in a sense of omnipotence, or grandiosity. In a
subgroup of borderline patients, to be discussed in the next section,
the fusion of self-concept, ego-ideal, and ideal object results in the nar
cissistic personality structure. However, in all borderline patients, the
tendency to feel omnipotent is strong, since the patients identify with
the all good self-image to protect against the painful all bad self-image
fraught with self-hatred and insecurity. The omnipotence defense also
protects against persecutory anxiety, since the patient lives in continual
fear of persecutory objects owing to the projection of oral rage.
Devaluation of the object is another major defense against persecu
tory anxiety, as the patient attempts to keep the object from appearing
dangerous. The aforementioned "primitive idealization" also serves
this function by endowing the object with magical qualities. The
patient appears to submit to a "magical," idealized figure but, accord
ing to Kernberg, treats the object ruthlessly, exploiting it only for pro
tection and gratification. When the object can no longer serve these
functions, it is easily devalued and rejected. However, devaluation is
not simply a response to the idealized object's failures in this regard: it
is a defense against the need for and fear of the object. Because the
patient still needs the protection of idealization, a cycle of idealization
and devaluation ensues. The patient's grandiosity fits both sides of the
cycle, as it is gratified by identification with a magical, idealized fig
ure and also allows for haughty devaluation of objects no longer
deemed necessary for protection or gratification.
From the brief discussion of the characteristic defenses of the bor
derline personality organization it can be seen that they all follow
from the predominance of splitting in the organization of the psyche.
The separation of all good and all bad self- and object images leads to
primitive idealization of the all good object image, fantasied omnipo
tence in the all good self-image, the use of denial to keep the oppo
sitely valenced images apart, projection and projective identification
to get rid of the all bad self-image, and devaluation to protect against
the persecutory object filled with projected aggressiveness.
The typical constellation of defenses seen in the borderline person
ality organization is found in the case of Miss B, who had great diffi
culty committing to treatment and had a history of failed relationships
(Kernberg, et al., 1988). This patient went to great lengths to be able to
begin treatment with her therapist, whom she considered the only
person who could help her. However, no sooner had she obtained
entry into treatment than she began to devalue him as "provincial,"
intellectually inadequate, lacking in sophistication, and without suffi
Otto Kernberg 205
cient self-assurance. At one point she seriously considered moving to

San Francisco to live with a man whom she believed to be far more
sophisticated than the therapist. In response to this devaluation,
which was delivered in a superficially friendly manner, the therapist
felt dejected, devalued, and despairing of the treatment until he real
ized that Miss B was wreaking her vengeance on the therapist for
what she presumed was his feeling of superiority when she was so
desperate to be accepted into therapy with him. Further, since her
mother adopted toward her the same attitude of quiet superiority and
subtle devaluation that made the patient feel stupid, inferior, despair
ing, and incapable of living up to expectations, the patient was evok
ing in the therapist the same sense of inadequacy and despair she had
been made to feel. Her image of the therapist as provincial, unattrac
tive, and intellectually slow was Miss B's self-image when she felt crit
icized by her mother. When the therapist pointed this out, "she now
reverted to a dependent relationship with the therapist, practically
without transition, while projecting the haughty, derogatory aspects
of herself as identified with mother onto the man from San Francisco"
(Kernberg et al., 1988).
Idealization was Miss B's initial defense until the relationship with
the therapist became a reality, at which point she shifted quickly to
devaluation to protect herself. This led to her primary defense of pro
jective identification: Miss B attempted to rid herself of her infantile,
negative self-image by projecting it into the therapist, whom she
attempted to control with subtle denigration. As she did so, she
became her aloof, haughty mother, adopting an omnipotent posture
while devaluing the therapist. When this dynamic was shown to her,
she quickly shifted roles once again, becoming dependent on the ther
apist—and presumably reverting to idealization to some degree, as
the man from San Francisco suddenly became the object of her deval
uation defense. Miss B showed no awareness of the shift nor of the
existence of conflict among her disparate ego states. Kernberg points
out that this quick oscillation and lack of awareness indicates the exis
tence of split object relations units, the pathognomonic indicator of the
borderline personality organization. This brief description of the case
of Miss B shows in condensed fashion many of the characteristic
defenses of the borderline personality organization. In the section on
treatment, we will have occasion to return to this case to demonstrate
the technique Kernberg advocates for the resolution of borderline
defensive operations.
The superego pathology of borderline patients is a direct product of
the defensive organization. Because the self-image is split between all
206
good and all bad, there is no opportunity for superego integration. As

we saw in the discussion of development, in Kernberg's view, normal
superego integration results from the integration of the sadistic,
highly fantastical superego forerunners with the ideal object-represen-
tations and with normal parental prohibitions as the split self- and
object-representations are integrated into the ego structure. However,
if the splitting process becomes entrenched as a defensive organiza
tion, instead of this integration into realistic self-regulating capacity,
the internal prohibitions of the psyche are left to the primitive, sadistic
self-attacks of the splitting stage. The ideal object-representations
become condensed into the sadistic superego forerunners, and no real
istic parental prohibitions are integrated into this primitive structure.
This is what happens in the borderline personality organization, as
Kernberg conceives it. Such patients can control themselves only with
harsh self-attacks, which explains their intensely negative self-devalu
ation. To defend themselves against the pain of these self-attacks, bor
derline patients identify with their ideal self-image and resort to
grandiosity, feeling themselves to be above normal human strictures.
The result is an oscillation between grandiosity and self-flagellation in
lieu of realistic values and prohibitions.
In Kernberg's's view, because the oral rage embedded in aggressive
object relationships is dissociated from libidinal object relationships
and not neutralized by libido, excessive oral aggressiveness is the fun
damental problem of the borderline patient. Again, one can see the
clear Kleinian influence in Kernberg's thinking. Kernberg believes that
in any given case there may be a constitutionally excessive aggressive
drive, or inability to tolerate frustration, extreme environmental frus
tration, or some combination of these issues. Whichever possibility
may apply in an individual case, the result is extreme oral rage that
must be managed by primitive defenses, thereby arresting further
development, including the genital organization of the libido. In one
case described by Kernberg (1975) the patient screamed so loudly at
her therapist during their interviews that she was heard throughout
the building. The therapist was so shaken by her rage that he some
times was "virtually trembling" after he saw her. On one occasion he
saw her by chance after a session and was shocked to see her relaxed
and smiling; the hospital personnel confirmed that she was relaxed
with other staff members. The intensity of the patient's rage split off
from other object relations is characteristic of the borderline personal
ity organization, according to Kernberg.
All these characteristics of the borderline personality organization
have direct application to the treatment process. However, before we
Otto Kernberg 207
proceed to draw out those clinical implications, we turn to Kernberg's

understanding of the narcissistic personality, the other form of severe
character pathology to which he directs a great deal of attention.
The Narcissistic Personality Organization
As we have seen, Kernberg (1976) traces the origin of the narcissistic

personality to either Stage 3 or Stage 4 of his developmental scheme.
He views the narcissistic personality as the use of a pathological
grandiose self to defend against splitting and other primitive defenses
of the underlying borderline personality organization. Kernberg's
acknowledgment that there is a subgroup of these patients who func
tion on an overt borderline level lends confusion to his seemingly
clear statement that the narcissistic personality is simply a more
superficially adapted borderline patient. Further, since the borderline
personality originates in Stage 3 of his developmental scheme, the
existence of some narcissistic personalities in Stage 4 would seem to
imply that not all have an underlying borderline personality organiza
tion. It is clear that Kernberg believes that pathological narcissism
encompasses a range of disturbances whose overt functioning ranges
from borderline organization to higher level character pathology.
At the highest level of narcissistic pathology are patients who have
unusual talents, skills, or attributes through which they derive abun
dant gratification from external sources. Such patients come for treat
ment because of serious neurotic symptoms or interpersonal difficulties
and have a good prognosis with psychoanalysis but may have diffi
culty engaging in treatment initially. The second subtype consists of
most narcissistic personality disorders; Kernberg believes these
patients also have a good prognosis with psychoanalysis proper. The
third subtype is the category of patients who exist on an "overt bor
derline" level. These patients are more chaotic than other narcissistic
personalities. They are distinguishable from other borderline patients
by their frequent outbursts of narcissistic rage and by their inability to
depend on others, in contrast to the excessive clinging of most border
line patients. A subgroup of this subtype includes patients who have
strong antisocial tendencies and hence a poorer prognosis. This sub
group shades into the fourth subtype: patients suffering from "malig
nant narcissism" who have virtually no structured superego and
consequently, have the poorest prognosis of all.
In Kernberg's (1975, 1984) view, narcissistic personalities are dif
ferentiated from other borderline patients by their fusion of the ideal
self, ideal object, and self-image, resulting in a grandiose self that
208
compensates for ego weakness so that the inability to manage tension

states is not readily apparent. This pathological construction allows
for better social adaptation than is characteristic of other borderline per
sonalities. Thus, the aforementioned nonspecific manifestations of ego
weakness are not seen in most narcissistic personalities. The degree of
success of this defense varies from well-integrated character pathology
to cases of overt borderline functioning in which the grandiose self
has not compensated well for the primitive defensive organization.
Because they have no realistic self-concept, these patients show an
exaggerated degree of self-reference. They need to be loved and
admired; to bolster the grandiose self, they actively seek continual
attention and admiration from the world. If they do not receive this
self-affirmation, they tend to feel bored and restless; the need for con
tinual bolstering by the environment is a telltale indicator that a pre
sumably positive self-concept is a defense against underlying
personality structure built upon splitting. The self-images of these
patients, like those of all borderlines, are split between all good and all
bad. They differ from other borderline patients in that the all good
self-image has become integrated as a stable structure. When the envi
ronment does not affirm the grandiose self, the negative devaluing
self-image appears. Consequently, grandiosity and self-devaluation
tend to alternate in consciousness, each split from the other.
This combination of grandiosity and need for affirmation is illus
trated by Kernberg's (1984) case of Mr. T, a social rehabilitation pro
fessional who entered analysis for help in his relations with women,
inability to empathize, and general boredom, irritability, and dissatis
faction with life. In the early phase of analysis Mr. T was finely
attuned to every move Kernberg made and was angered whenever a
detail of the material was forgotten. He complained that his girlfriend
was exploiting him, but he was dominant and exploitive in his rela
tionship with her, expecting her to guess his moods and respond to
his needs without his telling her what they were. Although he adopted
the posture of a helpless little boy exploited by powerful, aggressive
figures both outside the analysis (for example, with women) and
within it, Mr. T was, in fact, hostile and exploitive in both spheres. He
consistently denigrated Kernberg and summarily dismissed interpre
tations with which he disagreed—when he agreed with an interpreta
tion, he acted as though he already knew it. He frequently reacted to
interpretations by attacking Kernberg for trying to make him feel
guilty. When his response to interpretations was analyzed, he
expressed the belief that when Kernberg's comments were accurate,
they implied a "grandiose triumph" over him. Nonetheless, Kernberg
Otto Kernberg 209
frequently felt helpless, as though he were paralyzed by the patient.

During a protracted period of the treatment, the patient "confided"
Kernberg's shortcomings to people who were hostile toward Kern
berg and extracted information from them that he considered damag
ing to his analyst. At first he mentioned none of this in his analysis.
When he finally did admit to this pattern of behavior, he acknowl
edged a sense of power and excitement in feeling that he could con
trol and manipulate his analyst. This feeling turned out to be an
identification with his sadistic, controlling mother, the core of his
pathological grandiose self.
The case of Mr. T illustrates typically narcissistic split self-images;
he was either sadistic and ruthless, dominating others or the in
adequate, helpless little boy feeling exploited. Mr. T was more con
scious of the latter, but the analysis revealed that this negative
childhood self-image hid his grandiose identification with his sadis
tic mother. Once the grandiose self was revealed, the patient alter
nated between the two self-images, with Kernberg being the other
pole of the object relations unit. When the patient was the grandiose,
sadistic mother, Kernberg was the helpless little boy; when the
patient experienced himself as the latter, he felt that Kernberg was
the ruthless, exploitive mother.
The degree to which the environment denies these needs for contin
ual affirmation is the degree of difficulty the grandiose self has in
serving its function of effecting a stable organization that allows for
smooth surface adaptation. Thus, the stability of the grandiose-self
organization is much enhanced in narcissistic patients who possess
some type of special talent, skill, or attribute that brings narcissistic
supplies from the environment. The clinician can be surprised by the
depth of regression when such patients lose narcissistic bolstering
from the environment or regress within the treatment setting in
response to the analyst's failure to affirm their superiority. None
theless, Kernberg considers such patients to have the best prognosis of
all narcissistic patients.
Because of their continual need to "feed" the grandiose self, the
object relationships of these patients tend to be exploitive. Others are
valued for their ability to enhance the feeling of superiority and deval
ued when they do not serve this function. Consequently, there is little
warmth or depth to the personality or to interpersonal relationships.
The organization of the personality by the grandiose self dictates cold,
exploitive, even arrogant attitudes toward others. In the narcissistic
patient, the omnipotence defense becomes the organizing force of the
personality, with others being devalued because they represent a
Chapter 5
threat to the grandiose self. However, if objects enhance the grandiose

self, grandiosity will be projected onto them, resulting in idealization.
Kernberg is careful to differentiate the narcissistic personality from
other character pathology with narcissistic defenses. He believes that
the term narcissistic is overused because clinicians do not make the
distinction between narcissistic defense and the narcissistic personal
ity organization. Indeed, Kernberg points out that there are three lev
els of narcissistic defense. At the first level all defenses serve to protect
self-esteem to some degree; this characteristic of all neurosis is easily
distinguishable from the narcissistic personality organization. At the
second level are the many narcissistic defenses in character disorders
to protect and enhance self-esteem. For example, the hysteric may be
flirtatious and seductive to defend against penis envy. Her desire to
be thought of as beautiful and seductive serves to enhance her self
esteem, but the defense is directed against penis envy. When she is
not viewed as beautiful or her overtures are not responded to, she
does not crumble because her self-concept is not fused with her ideal
of being the beautiful seductress. Further, her relationships continue
to include warmth even when she is not responded to as she wishes to
be. At the third level is the narcissistic personality structure in which
libidinal investment in the self is equated with ideal qualities. A
woman with a narcissistic personality who desires to affirm her
grandiose self by being recognized as beautiful and seductive cannot
tolerate being viewed in any other way. When her overtures are not
responded to, she flies into a rage or attacks herself, and objects that
do not affirm the pathological grandiose self are devalued as worth
less. Consequently, the warmth and depth that are found in the
hysterical personality structure are absent.
There is a subgroup of narcissistic personalities for whom the
grandiose self does not allow for smooth social functioning. Such
patients display the same "non-specific manifestations of ego weak
ness" as other borderline patients, with the same resulting chaos and
displays of oral rage. In addition, these patients are narcissistically
demanding and are given to frequent outbursts of narcissistic rage
when their need for adulation is not met. Kernberg considers them a
risk for treatment, especially if they have antisocial tendencies.
Because the grandiose self is a condensation of the ego ideal with
the ideal self and the self-concept, narcissistic personalities tend to
have even less integration of early sadistic superego forerunners with
realistic parental prohibitions and the benign, loving ego-ideal and
ideal-object images than is true for most borderline patients. That is to
say, the ideal-self image tends to be absorbed into the pathological
211
grandiose self. Consequently, in lieu of some degree of realistic, inte

grated superego, some narcissistic personalities have an almost total
continuance of the early fantastical sadistic self-attacks. This primitive
superego forerunner tends to be projected onto objects, accounting for
the paranoia seen so frequently in these patients. Consequently, socio-
pathic tendencies are frequently stronger in narcissistic personalities
than in other borderline patients. According to Kernberg, sociopathy
lies at the severe end of the narcissistic continuum, making the degree
of sociopathy a major prognostic indicator for narcissistic patients.
Narcissistic personalities with some degree of concern for their effects
on others have a better prognosis than those who do not.
In its most extreme form, this process results in what Kernberg
(1984) calls "malignant narcissism." In these cases the idealized object
images that are normally integrated into the superego, are completely
absorbed into the pathological grandiose self. In Kernberg's (1984)
view, idealized object images are absorbed to some degree in all nar
cissistic personalities; however, in most cases, "a remnant of idealized
superego precursors remains outside the pathological grandiose self"
(p. 297), permitting some superego functioning. However, since no
such remnants exist in cases of malignant narcissism, ideal self-images
and ideal object images become a part of the grandiose self, with the
result that the sadistic superego forerunners express unmitigated
aggression and the grandiose self now contains all the aggression
within itself. Kernberg (1984) sum marizes: "The pathological
grandiose and sadistic self replaces the sadistic precursors of the
superego, absorbs all aggression, and transforms what would other
wise be sadistic superego components into an abnormal self structure
that then militates against the internalization of later, more realistic
superego components" (p. 298).
These unneutralized, highly sadistic, cruel, punitive self-attacks are
projected onto objects, resulting in a paranoid personality structure
as an alternative to self-directed sadism. As the patient can now
become the victim of humiliating and exploitive attacks by the object,
he or she will at times revert to sadism and attack objects to protect
against perceived persecution. This dynamic of projection and attack
is Kernberg's formulation of the frequent paranoid transference
regressions seen in cases of malignant narcissism.
Because the grandiose self is condensed with sadism, these patients
have a tendency to obtain sadistic glee from feeling victorious over
others, even when their own existence is at risk. Consequently,
such patients lie at the most severe end of the spectrum of narcissis
tic pathology and their prognosis tends to be poor. According to
212 Chapter 5
Kernberg, they are willing to self-destruct to feel the joy of triumph

over the analyst and to render the analyst impotent.
Kernberg is not definitive regarding the reasons for the pathologi
cal fusion of the ideal self, ideal object, and self concept in narcissistic
personalities. He attributes the formation of the grandiose self to
excessive oral aggressiveness, but cannot offer a decisive account of
why some borderline patients are able to construct such a narcissistic
defense, while others are not, but he does believe that certain familial
patterns are apparent in the histories of these patients. Kernberg's
(1975) clinical experience suggests that there is usually one chronically
cold parental figure who is indifferent and spitefully aggressive. This
cold aggressiveness leads to an increase of oral aggression in the child,
which exacerbates envy and hatred. The envy is defended against by
inciting envy in others. Again, the special quality or talent plays a key
role. Kernberg insists that such patients tend to have some realistic
attribute that does incite envy and admiration in others. Further, he
contends that this special quality was used by at least one parental fig
ure to achieve compensatory admiration. Consequently, the patient
tends to play a special role in the family, such as "the family genius,"
which serves to form a nucleus for the grandiose defense against oral
rage. Kernberg fully acknowledges that he cannot definitively explain
narcissistic pathology by this family pattern, but he does believe that
once the grandiose self is formed as the stable defensive configuration,
a cycle of self-admiration and devaluation of others takes place.
It is in his discussion of narcissistic pathology that Kernberg's
Kleinian influence is most apparent. He assigns a primary pathogenic
role to excessive aggressiveness and envy and the defenses against
them in the formation of the pathological grandiose self (Kernberg,
1975). In his view, the intensity of envy and hatred lead to an inability
to depend on others for fear of inciting oral rage and envy. Such
patients feel forced to adopt the posture of arrogant, callous indiffer
ence and haughty devaluation of others as if to proclaim their lack
of need for others. Consequently, such patients must spoil all they
receive from others to defend against envy of them. Any gratification
from another is an acknowledgment of dependence, an intolerable
feeling to the narcissistic patient. One can see in this analysis Klein's
(1975) formulation that the need to defend against dependence and
object contact is rooted in envy and defended by devaluation and
omnipotence. Even closer to Kernberg's view is Rosenfeld's (1971)
concept of the grandiose self as a pathologically aggressive structure
that defends against the libidinal, dependent self. The result of the
inability to receive from others is an empty, hungry self that envies
Otto Kernberg 213
others all the more. According to Kernberg, the empty, hungry,

enraged self is the deepest level of the self-concept of the narcissistic
personality disorder, but it may only appear at the end of treatment, if
the grandiose self defense has been successful.
These dynamics are illustrated in the case of Mr. T, discussed ear
lier, who either accepted interpretations as though he had already had
the insight himself or summarily dismissed them. Nonetheless, he
often used Kernberg's interpretations in his own work. Kernberg con
cluded that the patient was unable to depend on him for psychologi
cal exploration and instead "extracted interpretations" and put them
to his own use. Kernberg (1984) points out that when these reactions
were explored, "it emerged that Mr. T was protecting himself against
intense feelings of envy of me by utilizing for his own purposes what
ever he saw as new and good coming from me" (p. 214). The patient's
inability to depend on Kernberg was rooted in his need to defend
against envy, a need so strong that Mr. T was willing to deny himself
the value of his analysis to satisfy it.
Nonetheless, narcissistic patients will idealize others who they feel
possess what they lack. The idealization should not be mistaken for
genuine admiration and appreciation; rather it is a projection of the
grandiose self onto an object who is perceived to possess wished-for
qualities. Such objects are not, therefore, experienced as who they are;
they represent the patient's self. If such objects fail to serve the narcissis
tic purpose, by disappointing the patient or losing some admired qual
ity, they are easily given up and frequently devalued, reflecting the lack
of real attachment. Thus, idealization does not help narcissistic patients
achieve any genuine gratification from the object; even idealized rela
tionships leave them empty. Such patients have no capacity for empa
thy, and cannot experience depression, or mourn loss of objects; they
can only feel lack of narcissistic gratification.
Kernberg (1975) illustrates his view of the idealization typical of the
narcissistic patient in his description of the analysis of a professional
colleague who believed his analyst to possess perfect technique and
who gradually formed an image of him as an absolutely self-assured,
incorruptible, perfectionistic technician who was cold and distant but
masterful and reliable. It turned out that the patient was preparing to
switch to an analyst in another city in case he found any flaw in his
current analyst. The patient presented to his own patients the same
qualities he ascribed to his analyst, evincing a cold, distant, intellectu
ally precise approach of which he was quite proud. He became
extremely disappointed whenever the analyst did not conform to his
self-image; he felt threatened by the independence of the analyst. It is
214 Chapter 5
Kernberg's view that such an idealization showed neither concern for

nor real attachment to the analyst, as the patient expected to control
the analyst and was quite prepared to dismiss him summarily at the
first indication of a flaw. In addition, Kernberg points out that the ide
alization was a projection of the patient's grandiose self: the analyst
was seen as possessing the same characteristics as the patient and
no other qualities were tolerated. In Kernberg's view this exploitive
idealization is quite different from a genuine appreciation of admired
qualities in another.
Kernberg is careful to differentiate the idealization characteristic of
the narcissistic patient from other forms of idealization. The typical
borderline patient's idealization is of the primitive type discussed ear
lier, in which others, including the analyst, are seen as all good to pro
tect against contamination by the persecutory object, which is
dangerous by virtue of projected oral aggressiveness. In less primitive
character pathology, by contrast, idealization is a defense against
ambivalence and guilt over aggressiveness; thus genuinely loving
feelings are part of the relationship. At a still higher level of idealiza
tion, the object is viewed in accordance with higher level values and
superego functions. This type of idealization is characteristic of falling
in love. All three types of idealization differ from the projection of the
grandiose self, the idealization characteristic of narcissistic patients.
It should be noted that all pathological forms of idealization are
defensive constructions; they differ only according to the level of con
flict against which they defend. This view sets Kernberg's formulation
of idealization in clear opposition to Kohut's (1971) concept that ideal
ization in the narcissistic personality is a response to developmental
arrest (as will be discussed in chapter 6).
Kernberg's view of idealization in the narcissistic personality is an
application of Klein's concept of projective identification: the object
stands for the self and must therefore be controlled so as to fit the pro
jection. This is not a merger in Kernberg's view; the object must be like
the patient, not merged with the patient. Consistent with his develop
mental approach, Kernberg considers true merger to exist only in psy
chotic states. He feels that the terms merger and symbiosis are too
loosely used, and (as we will see in chapter 6) this is one of the major
points of disagreement between Kernberg and Kohut. In narcissistic
pathology, Kernberg believes that since self-object boundaries are
maintained, there is no true confusion between the experience of the
self and the object. The object is not fused with the self, it represents
the self and must conform to this representation. Because most people
do not fit the projection, interpersonal relations tend to fail, resulting
Otto Kernberg 215
in an exacerbation of emptiness that leads to a self-perpetuating cycle of

rage and envy, frustration, emptiness, more rage and envy, devaluation,
and a sense of impotence.
According to Kernberg, then, most patients with narcissistic pathol
ogy are accessible to psychoanalytic treatment but require intervention
unique to their pathology, as do other borderline patients. We now turn
to the specific type of psychoanalytic treatment Kernberg recommends
for both borderline patients and narcissistic personalities.
TREATMENT
Kernberg (1980) divides psychotherapy into three categories: unmodi

fied psychoanalysis, or psychoanalysis proper, psychoanalytic psy
chotherapy, which he calls expressive psychotherapy, and supportive
psychotherapy. As we have seen, Kernberg recommends psycho
analysis for the neurotic personality organization and for most narcis
sistic personalities and expressive psychotherapy for severe character
pathology, such as the borderline personality organization. He also
advocates psychoanalytic psychotherapy for milder neurotic cases in
which the unmodified psychoanalysis is not necessary. Interestingly,
Kernberg believes that for the latter group, supportive measures can
be included in the treatment because they do not confuse the patient.
For severe character pathology, however, inherent ego weakness ren
ders the patient too vulnerable to the gratifications of supportive tech
niques and confuses the patient about the role of the therapist and the
purpose of treatment. Further, since supportive therapy cannot
strengthen the ego structure, this form of treatment leaves the ego
deficient whereas expressive psychotherapy strengthens the weak ego
and can thereby result in appreciable improvement. In this sense
Kernberg opposes the orthodox view that the more severe the pathol
ogy, the more supportive the treatment should be. His contention is
that severe character pathology requires a psychoanalytic approach to
undo the primitive defenses and help the patient advance to higher
level ego structures.
Kernberg adopts Gill's (1954) tripartite definition of psychoanaly
sis: the position of technical neutrality, the use of interpretation, and
the systematic analysis of the transference. Accordingly, he defines
psychoanalytic psychotherapy as treatment that modifies one or more
of these principles. As discussed earlier, Kernberg (1980) views inter
pretation, clarification, and confrontation as the three primary psy
chotherapeutic techniques. Psychoanalysis proper uses predominantly
or exclusively interpretation, whereas psychoanalytic psychotherapy
216 Chapter 5
utilizes a mixture of interpretation, clarification, and confrontation.

Transference interpretation is used in psychoanalytic psychotherapy
as a primary tool but not as systematically as in psychoanalysis
proper; rather, the therapist selectively chooses to work on some resis
tances while supporting others. The therapist attempts to maintain
technical neutrality in psychoanalytical psychotherapy, but it is
expected that this will not always be possible with more severe psy
chopathology. The more the therapist is able to employ the three prin
ciples of Gill's tripartite definition of psychoanalysis, the more closely
the treatment approximates psychoanalysis proper; the more the
therapist must deviate from these principles, the more closely the
treatment approximates supportive psychotherapy. The latter utilizes
very little if any interpretation and relies instead on clarification,
abreaction, suggestion, and manipulation.
Treatment of the Neurotic Personality Organization
Although Kernberg (1980) has focused his work on the understanding

and treatment of severe psychopathology, he applies his object rela
tions theory to a general concept of psychoanalytic treatment that
embraces all levels of psychopathology. His object relations approach
to development and psychological structure leads him to reconceptu-
alize the therapeutic action of psychoanalysis. He points out that since
both drives and defenses are always expressed through object rela
tionships, intrapsychic conflict in neurotic cases involves object rela
tions units. Therefore, the psychoanalysis of neurosis involves the
breakdown of the ego and superego structures into their constituent
object relations components and the rebuilding of these structures
with a new integration of object relationships.
The therapeutic action of psychoanalysis in Kernberg's view is not
so much in making drive derivatives conscious as in making con
scious the object relationship building blocks of the psyche and reinte
grating them. Kernberg (1988) believes that his reconceptualization of
the psychoanalytic process helps the analyst gain insight into material
frequently obscured by the classical model. Since the psyche is formed
from the internalization of object relations units, not objects per se, the
patient forms an interpersonal relationship with the analyst based on
identification with either self or object, with the analyst perceived as
the other pole of the object relationship. In Kernberg's (1988) object
relations model the analyst is alert to the fact that patients will not
only enact themselves as their childhood selves, with the analyst as
the parental figure, but will also enact the parental figure, with the
Otto Kernberg 217
analyst as the childhood self. Often the analytic material appears not to
make sense because the patient is enacting an internalized object rela
tions unit with the roles reversed. Kernberg believes that this object
relations perspective on psychoanalysis broadens the therapeutic
armamentarium of the analyst.
For the neurotic personality organization, the analytic process
results in the breakdown of the ego and superego structures into
their constituent object relations units. As a result, the analysis leads
to the emergence of splitting and the rapid cycling of self- and
object images in a manner similar to treatment process with the bor
derline patient. At this point in the analysis of the neurotic person
ality, the treatment setting becomes the focus and sudden shifts in
the enacted self- and object images give a chaotic, confusing feel to
the analytic regression. Consequently, the analyst may feel tempted
to believe that the patient is not analyzable and yield analytic neu
trality. However, Kernberg points out that since the cause of the
chaos is the regression to splitting, the appropriate analytic inter
vention is the interpretation of the splitting. No extraanalytic mea
sures are necessary if the analyst is able to perceive the part object
relations units enacted in the analytic regression.
The best clue to understanding the enacted object relation is the
affective atmosphere generated by the patient and felt by the analyst.
Kernberg defines the countertransference as the totality of the thera
pist's affective responses to the patient, and he views countertransfer
ence reactions on a continuum from those that come mostly from the
therapist to responses that come primarily from the patient. Kernberg
believes, like Winnicott (see chapter 4), that countertransference reac
tions should be neither repressed nor split off, but used as a source of
information about the patient. In general, Kernberg feels that the more
severely character disordered patients and neurotic patients in
moments of analytic regression tend to elicit more intense emotional
reactions from therapists than do patients with higher level disorders.
Nonetheless, even in the more routine periods of the analysis of neu
rotic disorders, the countertransference can be a valuable clue to
understanding the currently enacted object relationship.
While countertransference is useful during any part of an analysis,
it is Kernberg's contention that the countertransference is an espe
cially good indicator of analytic regression. Chaotic, intense, quickly
oscillating countertransference reactions are likely to indicate a regres
sion in the patient that may go unnoticed if the countertransference
is not attended to. By regarding the countertransference in this
way, Kernberg is applying Racker's (1968) notion of "concordant"
218 Chapter 5
countertransference to the treatment of the neurotic patient. (Here

again one sees the impact of the Kleinian movement on Kernberg's
thinking.) He finds projective identification and especially Racker's
concepts of complementary and concordant countertransference to be
of special importance in the treatment of these patients and suggests
that the analyst pay close attention to their affective responses to the
patient and rather than act on them, use their understanding of the
enacted object relationship to inform their interpretations.
One can see from this approach to analytic regression that Kern
berg opposes the view that analysts need at certain points in the treat
ment to offer themselves as a "real" person or to become more
"human," an approach that considers the therapeutic action of analy
sis to lie in the patient's internalization of the analyst's "maternal
function." Kernberg (1980) is critical of Fairbairn and Guntrip for
adopting this approach. Kernberg makes a clear distinction between
maternal empathy and analytic empathy: the latter includes the ana
lyst's understanding of the patient's dissociated and repressed mater
ial, acting out, and nonverbal interaction with the analyst and thus
involves a great deal more than the maternal empathy. Kernberg cor
respondingly feels that identification of the therapeutic action of psy
choanalysis with the internalization of the mothering function is
misleading and misses the unique value of psychoanalysis. He
believes that empathy is a crucial ingredient of the analytic process—
but as a prerequisite of interpretation, not a substitute for it. The ana
lytic resolution of conflict is served by interpretation of the dynamic
unconscious and by the integration of the revealed part-object rela
tions units that compose the personality, not by the offering of a
personal relationship that violates technical neutrality.
Kernberg's concept of the analytic treatment of the neurotic per
sonality organization is an application of his object relations model
of personality structure and psychopathology within the context of
the three cardinal principles of psychoanalysis. Unlike Fairbairn,
Guntrip, and Winnicott, Kernberg does not believe that the princi
ples of psychoanalysis need to be modified, but he does believe that
they must be applied in a manner somewhat different from that sug
gested by the classical model of psychoanalysis. Kernberg's model
focuses on making conscious object relationships that are enacted in
the transference paradigm and on using of countertransference as
a primary tool to achieve understanding, rather than on making
conscious drive derivatives through defense analysis. Thus, his con
tribution to psychoanalysis is both a reconceptualization of the clas
sical view of the aims of the analytic process and a counterpoint to
Otto Kernberg 219
the radical departures in analytic technique suggested by other

object relations interpretations of psychoanalysis*
Kernberg's (1980) approach to the analytic treatment of a neurotic
case can be seen in the psychoanalysis of a professional man in his
forties who entered analysis for help with his chronic marital con
flicts, severe work inhibition, and occasional sexual impotence. The
patient was potent with prostitutes and achieved sadistic sexual grat
ification with them. He feared and depreciated his father, who was a
prominent man. His mother submitted to the father but was con
stantly complaining, guilt-evoking, and hypochondriacal. The patient
presented in the analysis as shamefully submitting to his wife who
imposed restrictions on him and his work. He devalued Kernberg
and kept him analytically impotent with his nagging protests about
treatment and his amused reactions to interpretations that seemed as
if they were prods accompanied by the message, "You can do better
than that." When Kernberg realized how impotent he felt, he inter
preted to the patient that he was enacting his nagging wife while pro
jecting into Kernberg his self-image of impotence and helplessness.
The patient then began to view Kernberg as forceful, as someone who
would not put up with a wife's constant complaints, and, indeed,
Kernberg felt a sudden sense of power. This was understood as an
enactment by the patient with Kernberg as his powerful, brutal father
forcing him into homosexual submission. Kernberg went on to point
out that the patient's search for prostitutes and his fear of competing
with the analyst/father lest he lose his masculinity were defenses
against the temptation to submit. The patient responded to these
interpretations by becoming depressed over opportunities lost
because of his neurotic fears. He realized for the first time that part of
the reason he and his wife did not have children was his inability to
assert himself against her reluctance. In response to this painful real
ization, Kernberg felt a strong sense of empathy and positive concern
for the patient.
This brief vignette from the analysis of a neurotic patient involved
three phases of object relations enactments. In the first phase the ther
apist was the projected self-representation of the patient, impotent in
response to the sadistic mother/wife. In the second phase the thera
pist was the powerful, brutal father and the patient was the helpless,
fearful child. In the third phase Kernberg felt empathy for the
patient's self experience. Thus the analysis involved the continual
interpretation of enacted object relationships. Further, Kernberg's abil
ity to understand the object relationship enactments was critical to
his awareness of his changing countertransference feelings. This
220 Chapter 5
vignette, although brief, illustrates Kernberg's reconceptualization of

the therapeutic action of psychoanalysis as the uncovering of internal
ized object relations units by the understanding of transference-
countertransference interactions.
Despite his reconceptualization of the psychoanalytic process,
Kernberg's primary efforts have been focused not on the psychoanaly
sis of neurotic personalities but on expressive psychotherapy for bor
derline and narcissistic personalities, and it is to his contributions in
this area that we now turn.
Treatment of the Borderline Personality Organization
Kernberg's recommendations for the psychotherapy of the borderline

personality organization follow closely his conceptualization of the
pathology. Because the fundamental pathogenic issue is excessive oral
aggressiveness split off from libidinal object relations, the critical
transference development is excessive oral aggressiveness directed to
the therapist. The negative transference tends to become the treatment
focus as the early pathogenic object relationships are activated.
Patients project their intense aggressiveness onto the therapist, who
quickly comes to represent a hated figure of the past. Because the
intense oral aggression interferes with the therapeutic alliance, both in
its direct and projected forms, it must be addressed immediately. It is
one of Kernberg's cardinal technical principles that the negative trans
ference must be interpreted quickly and forcefully with the borderline
patient. If the therapist chooses to ignore the negative transference,
there is no chance to disrupt the projective-introjective cycles and no
opportunity to build the therapeutic alliance, a prerequisite for the
observing ego. In Kernberg's view, the consistent interpretation of the
negative transference is necessary for the development of the patient's
observing ego, and the enhancement of the observing ego helps in the
disruption of the projective-introjective cycles.
While Kernberg advocates early, forceful interpretation of the
negative transference to build the observing ego, he recognizes that
some degree of preexisting alliance must be present for transference
interpretations to be effective. He resolves this problem by pointing
out that there are some modulated, positive feelings toward the
therapist that are not part of the idealizing defense. One of his treat
ment principles for the borderline patient is that these feelings
should not be interpreted because they aid in the development of
the therapeutic alliance. Thus, while treatment involves confronta
tion of the negative transference, the more "modulated" aspects of
Otto Kernberg 221
the positive transference are not interpreted but utilized to advance

the therapeutic process.
Interpretation of the negative transference may seem obvious when
it is acted out, but in some cases the negative transference is not
clearly seen. Kernberg distinguishes between the manifest and latent
negative transference in the borderline patient. In the latter case the
patient maintains a presumably friendly, albeit superficial, attitude
toward the therapist, devoid of signs of aggressiveness, while splitting
off his or her negative object relations from the treatment situation. If
the therapist accepts the patient's superficial, detached presentation,
the relationship comes to be based on a denial of the negative transfer
ence, a denial that fosters the fundamental pathology. The therapeutic
intervention is to interpret the split-off negative transference in order
to bring the patient's aggression into the analytic relationship.
Kernberg fully acknowledges in cases of this type that therapeutic
interventions generate anxiety. However, he feels this anxiety indi
cates therapeutic progress since it reflects the entry of aggressiveness
into the transference and allows for the resolution of splitting.
Kernberg's (1975) approach to splitting in borderline patients who
maintain a detached, superficial attitude to the therapist is illustrated
in the case of a female patient suffering from drug and alcohol addic
tion. The patient made considerable progress in the hospital and
seemed to make a good adjustment after discharge, but her attitude in
outpatient psychotherapy was shallow and "conventionally friendly."
After some months of sobriety, she became drunk, depressed, and sui
cidal and required rehospitalization. She kept this episode from her
therapist, who found out about it only after she was rehospitalized.
After discharge, she denied all transference and emotional implica
tions of the episode, despite her memory of anger and depression dur
ing it. At this point the therapist began a long effort to correlate the
patient's detached attitude in treatment with her alcoholic crisis. After
two more episodes of this type it became clear that "she was experi
encing the therapist as the cold, distant, hostile father who had
refused to rescue her from an even more rejecting, aggressive mother"
(p. 95). Kernberg summarizes the situation this way:
The patient felt that if she really expressed to the psychotherapist-father

how much she needed him and loved him, she would destroy him with
the intensity of her anger over having been frustrated so much for so
long. The solution was to keep what she felt was the best possible rela
tionship of detached friendliness with the therapist, while splitting
off her search for love . . . and her protest against father in alcoholic
222 Chapter 5
episodes during which rage and depression were completely dissociated

emotionally from both the therapist and her boyfriends [p. 95].
Kernberg's central point is that it was necessary for the therapist to

introduce all this material into the transference through a systematic
confrontation with the patient's effort to split off her aggressiveness
toward the therapist. This type of intervention—confronting of the
"friendly" relationship toward the therapist with the negative trans
ference—aims at the "undoing" of splitting. Kernberg acknowledges
that in this case the therapist's interventions increased the patient's
anxiety to the point that she became even more distrustful and angry,
reverting to her old alcoholic patterns; he acknowledges further that
interpretation did not suffice to control this behavior and that there
was another hospitalization. Nonetheless, he contends that although
the patient appears to have done worse "from a superficial point of
view," the therapist felt that "for the first time he was dealing with a
'real' person" (p. 96).
Object relations are split into part self- and part object image units,
and the primary goal of the treatment of the borderline patient is to
integrate these part object relations units into whole self- and whole
object images. To accomplish this task the therapist must identify and
label the split-off self- and object images. For example, when Miss B,
described earlier, devalued her therapist, she was enacting her
mother's sadistic attitude toward her and evoking in the therapist the
feelings she had experienced in response to her mother's criticisms.
Her other primary object relations enactment was of a helpless child
dependent on and submissive to the mother. In fulfilling the therapeu
tic task of identifying the roles patients enact for both themselves and
their therapists and then pointing out the conflict between them, Miss
B's therapist interpreted her haughty devaluation as an enactment of
herself as her mother while projecting into the therapist the feelings
she had when her mother criticized her. When Miss B immediately
became dependent on the therapist and projected her haughty
grandiosity onto her boyfriend, the therapist interpreted this new
object relations constellation. In Kernberg's view, the therapist must
continually search for the object relations unit enacted at each phase of
the treatment, label for the patient the self- and object images enacted,
and identify the connecting affective link. Additionally, the therapist
must point out the existence of contradictory self- and object images
and their connection in order to bring them together into integrated
whole object representations.
This case illustrates the primary role of the countertransference in
Otto Kernberg 223
the understanding of the enacted object relationships in the treatment

of the borderline patient. The therapist understood Miss B's enact
ment of the haughty mother and the attacked, helpless child object
relationship when he realized that he was feeling the victimized help
lessness that the patient felt in response to her mother's criticisms.
This countertransference response was the primary clue in unlocking
the mystery of the interaction, and Kernberg believes this is typically
the case with borderline patients because of the dominant use of pro
jective identification. That is, the patient tends to communicate by
projecting feelings into the therapist, rather than by verbalizing. In
formulating transference interpretations to borderline patients, the
first task for therapists is to sort out and understand their own affec
tive responses. Only after therapists have differentiated their affective
responses to the patient are they in a position to identify the actors in
the drama that is being enacted with the patient.
This task is rendered difficult by the fact that the patient shifts so
rapidly between different object relations units. The rapid reversal of
self- and object images is illustrated by events in a therapy session of
an 18-year-old girl who initially spoke in a deeply emotional manner
about matters that concerned her very much and then suddenly
became bland and indifferent regarding these same issues (Kernberg,
1975). Kernberg tried repeatedly to explore this material with the
patient but eventually gave up in frustration. When he ceased his
active effort, the patient suddenly became insistent that he give her
advice about a reality situation. As the patient became more demand
ing, Kernberg felt distant and impatient with her. He then realized
that in the first part of the session she had enacted her mother as a dis
tant, aloof, cold person while he was her self-image as a demanding,
frustrated child. In the later part of the session Kernberg had become
the rejecting mother and the patient was the needy, demanding, frus
trated child. When Kernberg realized these mother-child roles were
being played out, he was able to formulate an interpretation. It needs
to be underscored that Kernberg was able to grasp the meaning of the
patient's sudden behavioral shift when he realized that he felt distant
and impatient with her and was subsequently able to connect these
feelings with the behavior of her indifferent mother. In Kernberg's
view, by paying close attention to their emotional reactions to
patients, therapists will frequently be able to link the borderline
patient's presumably impulsive, chaotic behavior with a crucial
self-object unit from the patient's past.
Once the therapist has been able to interpret the currently operating
object relationship, special attention is paid to the patient's response
224 Chapter 5
to this interpretation. Kernberg believes the patient's reaction to the

awareness of the enacted object relationship will frequently reveal an
additional component of the object relationship structure of the per
sonality. For example, when the therapist interpreted to Miss B that
she was enacting her mother while putting her therapist in the posi
tion of her childhood self, she responded by reverting to her previously
dependent relationship on the therapist and projecting her haughty,
sadistic object image onto her boyfriend. In Kernberg's view, a frequent
response to the interpretation of an object relationship is role reversal of
the same object relationship. However, other responses are possible,
including the intensification of the same object relationship with the
same role enactment. If the patient's response to the interpretation
of the object relationship unit reflects a different aspect of the object
relationship structure of the patient, an interpretation of this newly
enacted object relationship should be made.
Eventually, the patient's response to the therapist's insistence on
interpreting splitting is to use one or more of the primitive defenses
characteristic of the borderline personality organization. As we have
seen, all these defenses are designed to foster the splitting process.
When the dominant use of splitting is threatened by the therapist's
systematic interpretations, the patient resorts to other primitive
defenses—projection, projective identification, primitive idealization,
omnipotence, denial, and devaluation—in a desperate effort to keep
apart the good and bad self- and object images. The therapeutic task is
to interpret these defenses as systematically as splitting is interpreted.
Only when these defenses are resolved is the patient able to relinquish
splitting as the primary organizing principle of the psyche.
Kernberg advocates limit setting, clarification of ego and object
boundaries, and confronting patients with their own aggressiveness
when they direct their excessive oral aggressiveness at the therapist.
Similarly, he believes primitive idealization should be treated by con
fronting patients with their unrealistically exaggerated positive feelings
and inquiring as to their origin. The point of the investigation is to show
patients that their idealization defends against their persecutory fears
of the therapist, which in turn are a projection of their oral aggressive
ness. Kernberg acknowledges that patients will attempt to maintain
their idealized perception of the therapist and that the unrealistic nature
of their transference perception must be confronted over and over
again. Only when the idealization is pierced will patients' paranoid
fears and primitive aggressiveness become manifest.
Denial, omnipotence, and devaluation are all addressed similarly.
In Kernberg's view, denial can be directed against sectors of subjective
Otto Kernberg 225
or external reality. In either case patients are confronted with the real
ity of the oral aggressiveness that they prefer to deny. Omnipotence
and devaluation are often not apparent, and the therapist interprets
these attitudes to the patient to make conscious the use of the defenses
before confronting the reasons for their existence. All these defenses
must be systematically undone for the ego to become integrated and
develop higher level defenses. This undoing requires a combination of
clarification, confrontation, interpretation, and limit setting.
A good illustration of Kernberg's approach to the defenses of the
borderline patient is provided by his account of the treatment of an
obese patient who believed that she had a right to eat whatever she
pleased and that she was still entitled to be "admired, pampered, and
loved" (Kernberg, 1975, p. 102). She felt entitled to come for therapy at
any time she wished, and to behave as she chose, including leaving
cigarette ashes all over the furniture in the therapist's office. "It was
only after the therapist made very clear to her that there were definite
limits to what he would tolerate, that she became quite angry,
expressing more openly the derogatory thoughts about the therapist
that complemented her own feelings of greatness" (p. 102). The
patient's conscious inferiority feelings had until this point masked her
omnipotence. The therapist's confrontation of the patient's entitled
behavior made conscious her feelings of greatness, which had been
operating as an omnipotent defense. That type of intervention
Kernberg believes, must be consistently applied to "undo" the
defense. While Kernberg does not say what was defended against in
his reporting of this vignette, he makes clear that omnipotence is char
acteristically used against awareness of oral aggressiveness and
threatening dependency needs.
To summarize, Kernberg believes that for the integration of split
object relationships his treatment principles gain expression in a series
of therapeutic steps (Kernberg et al., 1988). Because of the chaos of
rapidly shifting self- and object image enactments, the first step for
therapists is the exploration of their countertransference. Once thera
pists understand the role or roles they are being given to enact by
the patient, they are able to identify the self- and object images
enacted in the transference. In the third step, the therapist interprets
the transference paradigm by "naming" these split object relations.
Fourth, the therapist pays special attention to the patient's response
to the interpretation of the object relations enactment and interprets
this response if it involves a shift to a new dyad. Finally, the inter
pretation of splitting will eventually result in the appearance of one
or more of the primitive defenses characteristic of the borderline
226 Chapter 5
personality, and this too must be interpreted systematically to

resolve splitting definitively.
The integration of split object relations units in a borderline
patient treated by expressive psychotherapy is illustrated in the case
of Miss L, a Latin American artist in her twenties, who sought help
for a severe sexual inhibition, although her diagnosis was borderline
personality organization with severe depression and schizoid ten
dencies (Kernberg, 1984). Miss L's sexual fantasies involved the
mutilation of her and her partner's genitals during sexual inter
course. Her initial transference paradigm was organized around her
desire for the therapist to rape and kill her during sexual inter
course. Her second major transference involved a view of herself as
a dependent child, with the therapist seen as a motherly father from
whom she could receive warmth, protection, and love if she could
suck his penis. It became clear to Kernberg that Miss L's sexual
inhibition was a result of her inability to separate the two trans
ference constellations, an inability that lead to her fear that her
hatred and love would come together and that she would destroy
her only potential source of love, warmth, and protection. Later in
treatment Miss L's fear of orgasm was connected to a fear of "uncon
trollable wetness" and a fear that her personality would dissolve
into impersonal fragments. Kernberg summarizes the situation:
The predominance of splitting mechanisms, the fear of conflicts related

to severe oral frustration, and the regressive dangers of the oedipal situ
ation all blocked sexual excitement and orgasm. At a still later time
Miss L was able to fantasize more elaborate sexual experiences with
men, the therapist in particular, which centered upon letting herself go
and urinating during orgasm and which expressed her longings for
dependency and sexual gratification in more synthetic ways [p. 109].
Kernberg notes that after this development Miss L began to date more
and engaged in petting. However, a negative therapeutic reaction
ensued when Miss L became aware of a primitive, sadistic superego that
prohibited further improvement in her relations with men. She now sub
mitted to this primitive superego which appeared to be a condensation
of the "hated and hateful pregenital mother with the feared oedipal
rival" (p. 109). After working through this primitive superego, Miss L
was able to establish a sexual relationship with an appropriate object.
In Kernberg's view, this case shows the resolution of pathology in a
borderline personality organization by the integration of the two pri
mary split object relations units enacted in the transference. The major
Otto Kernberg 227
therapeutic shift occurred when Miss L became aware that her hatred,
death wishes, and fear of destruction involved the same person
toward whom she looked for the fulfillment of her needs for love,
warmth, and protection. The recognition that she had both sets of feel
ings toward the same man resulted in the integration of her view of
men and the capacity to open herself to her genital longings. After her
dependency longings became manifest, Miss L began to describe in
more detail her sexual fantasies and her tolerance of genital longings
increased. At that point Miss L began to have sexual experiences with
men; concurrently she became more productive at work.
The persistent search for the object relations role enactments is to be
found in the case of Miss N, who presented with a borderline person
ality organization with obsessive and schizoid features (Kernberg,
1984). In the first period of treatment, the transference paradigm
involved a masochistic search for a warm, giving, but sadistically
powerful father. Then Miss N shifted to a preoedipal mother transfer
ence and accused the therapist of being cold and rejecting. This para
digm alternated with her view of Kernberg as a sexually exciting,
powerful, dangerous man. When her fear of her sexual longings for
the therapist/father were interpreted, she regressed markedly. Miss N
let the therapist know that he must say only "perfect and precise
things that would immediately and clearly reflect how she was feeling
and would reassure her [he] was with her" (Kernberg, 1984, p. 129).
(As we will see in chapter 6, Kohut, 1971, referred to this need as the
"mirror transference.") During this period Kernberg only pointed out
to the patient how frightened she was of his overpowering her with
his comments and that he understood her need for him to understand
without her having to tell him. In response to this limited interpretive
stance, the patient appreciably improved, but when Kernberg
attempted to explore the two types of transference, no progress was
made. Eventually, Kernberg interpreted that Miss N had two alternat
ing views of him: in one he was a warm, receptive mother, in the other
the sexually tempting, dangerous father figure. Miss N then revealed
that when Kernberg was active she saw him as "harsh, masculine, and
invasive." When he listened more passively, he was "soft, feminine,
and somewhat depressed." After Kernberg interpreted this transfer
ence paradigm as her effort to avoid the conflict between her need for
a nurturing mother who forbade sex with the father and the need to
be a receptive woman to a masculine man, the treatment continued to
advance toward oedipal conflicts.
This case demonstrates Kernberg's technique of integrating object
relations in the transference. Even though at a major point in the
228 Chapter 5
treatment Kernberg felt compelled not to interpret actively, he eventu

ally was able to interpret Miss N's need to place him in the passive
role he had been playing in that phase of the treatment. A cardinal
tenet of Kernberg's approach to the psychotherapy of the borderline
patient is that the therapist must continually reflect on the role enact
ments between patient and therapist to understand the interaction,
especially when the treatment appears not to be progressing. Miss N's
putting Kernberg in a passive role was an enactment of a crucial com
ponent of the treatment, not an interference with it. When Kernberg
discerned what role he was playing and interpreted that to the
patient, the treatment was able to progress.
Despite Kernberg's adherence to interpretive principles, he believes
that interpretation is almost never sufficient for most borderline
patients. As we have seen, Kernberg believes that the borderline
patients lose self-object boundaries when the projective-introjective
cycles become intense and quickly alternate. Consequently, when the
transference intensifies, it may become psychotic, but the reality sense
is maintained in other situations because the patient has fundamen
tally intact self-object boundaries. This is Kernberg's concept of the
transference psychosis. To resolve it, parameters must be used. The
therapist clarifies the distinction between the patient and himself or
herself and limits the acting out of the negative transference. The char
acteristic development of transference psychosis is another reason for
the employment of noninterpretive intervention and contraindicates
psychoanalysis proper.
The addictive case referred to earlier illustrates the use of parame
ters in Kernberg's treatment approach to the borderline patient. The
intense negative transference was interpreted, but it was also limited
by the use of the hospital. This is a good illustration of why Kernberg
does not believe in psychoanalysis proper for these patients. The act
ing out of the negative transference tends to become so extreme that
extra-interpretive means are frequently necessary to control it so that
the treatment process can take place. Even in the case of patients who
act out the negative transference directly in the treatment setting, the
therapist must often set limits on what will be tolerated. This is not
simply protection for the patient and therapist. In Kernberg's view,
the limit setting is necessary for the development of the observing ego
and helps the patient reestablish reality boundaries. Limit setting
extends to the acting out of the patient's intense oral aggressiveness,
which, in Kernberg's view, is "instinctually gratifying." (Here
Kernberg applies his belief in the drive component of his object rela
tions model directly to his technical approach. Since he views aggres
Otto Kernberg 229
siveness as a drive, it follows that the persistent oral rage attacks of

borderline patients, if allowed to continue, provide drive gratifica
tion.) From Kernberg's point of view, as long as the patient's repet
itive behavior brings instinctual gratification, working through is
not possible; thus, the setting of parameters becomes critical to the
treatment process.
It should be noted that all interpretations cited that were aimed at
resolving oral aggressiveness and undoing defenses have focused on
the here-and-now transference. Kernberg believes that genetic con
structions should be avoided during the greatest part of the treatment
of the borderline patient. Owing to the fragility of the ego organiza
tion of such patients, such interventions tend to foster confusion
between past and present, fantasy and reality. Since the purpose of the
treatment is to help secure the patient's distinction between percep
tion and reality, Kernberg believes interpretations should be focused
on the patient's interpretation of here-and-now reality. According to
Kernberg, during most of the treatment the origins of the patient's dis
torted perceptions should be included in the intervention only when
aspects of the past are conscious; in this case, the connection of the
patient's transference perceptions with past history can help the
patient make the distinction between reality and perception. Genetic
reconstructions in the usual sense of making conscious the uncon
scious past are recommended only after the reality sense is more fully
developed near the conclusion of treatment.
By way of summary, one can see that Kernberg elucidates nine car
dinal principles for the treatment of all borderline patients, whatever
the particular character disorder may be. First, the negative transfer
ence must be made a major treatment focus, whether its manifestation
is latent or manifest. Second, the therapeutic alliance/observing ego
must be built via a combination of negative transference analysis and
noninterpretation of the modulated positive aspects of the patient's
relationship to the therapist. Third, and most important, the split
object relations units must be systematically integrated by identifying
and labeling them for the patient and then pointing out the conflict
between them. Fourth, special attention must be paid to the counter
transference for the identification of these split object relations units.
Fifth, attention is focused on the patient's responses to interpretations
of the enacted object relationships. Sixth, all related defenses must be
systematically "undone" by persistent interpretation. Seventh, to
resolve the acting out of the negative transference, and to help foster
reality testing and the observing ego, parameters must be employed to
prevent "instinctual gratification" of aggressiveness in the treatment
230 Chapter 5
sessions. Eighth, the transference psychosis must be managed by clari

fication of reality as well as by parameters to reestablish self-object
boundaries. Ninth, the transference analysis must be limited to here-
and-now interpretations, with genetic reconstructions left to the end
of treatment, if they are used at all.
These principles define Kernberg's approach, built on the work of
Klein and ego psychology, to the treatment of the borderline patient.
Kernberg's treatment recommendations are a consistent application of
his belief that severe character pathology should be treated by modi
fied psychoanalytic psychotherapy. However, there is one subgroup
of borderline patients who he feels can manage and benefit from
unmodified psychoanalysis: the narcissistic subgroup of borderline
patients.
Treatment of the Narcissistic Personality Organization
Kernberg takes the view that because narcissistic personalities are able
to utilize the grandiose-self defense, the danger of regression and psy
chosis and therefore the need for parameters, is not as great, as with
other borderline patients. On the other hand, because the grandiose
self appears to be so effective, it is even more difficult to resolve than
other borderline defenses. For both these reasons, Kernberg (1975,
1984) advocates unmodified psychoanalysis, the aim of which is the
revelation and undoing of the grandiose self. For the subgroup of nar
cissistic personalities who present overt borderline personality organi
zation, psychoanalytic psychotherapy is the treatment of choice, as it
is for most borderline patients.
The grandiose self is a stubborn, effective defense and must be con
sistently interpreted with a view to revealing the oral rage and perse
cutory anxiety that lie beneath it. As discussed earlier, the grandiose
self will be projected onto the analyst, and this idealization is also to be
treated as a defense against rage and the paranoid fear of the therapist.
Critical to Kernberg's treatment recommendations is the idea that
the grandiose self and idealization that must be undone by inter
pretation. As we will see in chapter 6, Kernberg criticizes Kohut for
recommending the acceptance of idealization without interpretation.
To Kernberg, this is supportive psychotherapy because it bolsters
defenses rather than undoing them; for him a psychoanalytic approach
must be directed at the unearthing and resolution of the oral aggressive
ness, persecutory anxiety, hunger for objects, and fear of dependence
that underlie the grandiose-self and idealization defenses.
According to Kernberg, narcissistic patients present other special
Otto Kernberg 231
clinical problems. Their excessive inability to depend on others, even

when idealization is prominent, result in special resistances. Nar
cissistic patients have more difficulty than other borderline patients in
receiving help, since help tends to elicit envy. The analyst's interpreta
tions tend to be devalued or ignored, rather than used for insight.
Consequently, they tend to spend countless sessions, sometimes for
prolonged periods, either overtly devaluing the analyst or trivializing
the process by keeping it superficial and giving it little importance.
Kernberg's approach to these narcissistic resistances is to interpret
them as defenses against envy, rage, and the anxiety of dependence.
Although Kernberg acknowledges that the resolution of these resis
tances takes a long time, he is opposed to allowing them to be enacted
without interpretations, as he feels that only a systematic interpretive
stance can eventually lead to their undoing.
Kernberg's treatment of narcissistic resistances is illustrated in his
discussion of a patient who fell in love with a woman whom he ideal
ized as beautiful, warm, and gifted. After they married, he became
bored with and indifferent toward her. In the analysis he became
aware that he treated his analyst in a similar manner and depreciated
all he offered in order to defend against his hatred and envy of him.
Subsequently, he became aware of his hatred of his wife for possess
ing all he lacked. After this awareness, he responded to his wife's
expressions of love for the first time. "His awareness of his aggressive
disqualification of her and his analyst, and his increasing ability to tol
erate his hatred without having to defend against it by destroying his
awareness of other people, made both his wife and his analyst 'come
alive' as real people" (Kernberg, 1975, pp. 237-238). This vignette
demonstrates Kernberg's approach of interpreting the narcissistic
patient's devaluation of the analyst and resistance to receiving help as
a defense against envy and aggressiveness.
The use of the grandiose self as a defense against receiving help
from the analyst constitutes the primary narcissistic resistance and
tends to become the focal point of the treatment. Because the grandiose
self is so well organized, even if the patient responds initially to the
analyst's interpretation of the resistance, he or she will quickly revert
to the defense. In Kernberg's view, the treatment process tends to
become a repetitive cycle involving the analyst's interpretations of the
patient's resistances, the patient's response and subsequent resurrec
tion of the narcissistic defense, and the analyst's interpretation of the
renewed defense. The following case illustrates this focal conflict in the
treatment of the narcissistic patient (Kernberg, 1975).
The patient complained ceaselessly of monotony and boredom in
Chapter 5
the sessions and insisted that treatment was hopeless; meanwhile, his
outside life seemed to be going well. Kernberg pointed out that to the
patient his criticism of analysis was an indirect devaluation of
Kernberg as the provider of useless treatment. While the patient ini
tially denied this, he later admitted to Kernberg that he blamed him
for the failure of the analysis and was surprised to find himself so
pleased to continue his treatment with him. At that point Kernberg
was able to interpret that the patient had, in fact, been quite satisfied
to view him as worthless while achieving success in life. In response
to this interpretation the patient became extremely anxious, feared
that Kernberg hated him, and developed paranoid fantasies of
Kernberg wreaking vengeance upon him. Kernberg interpreted that it
was precisely this fear of attack that lay behind the patient's need to
reassure himself that he was not in analysis by repeatedly proclaiming
that nothing of significance was occurring in the sessions. The patient
expressed admiration that Kernberg had not been derailed by his con
tinual assertions that analysis was a failure. Kernberg (1975) then
describes the crux of the ensuing process this way:
At the next moment, however, he thought that I was very clever, and that
I knew how to use "typical analytic tricks" to keep 'one up' over patients.
He then thought that he himself would try to use a similar technique with
people who might try to depreciate him. I then pointed out that as soon
as he received a "good" interpretation, and found himself helped, he also
felt guilty over his attacks on me, and then again envious of my "good
ness." Therefore, he had to "steal" my interpretations for his own use
with others, devaluating me in the process, in order to avoid acknowledg
ing that I had anything good left as well as to avoid the obligation of feel
ing grateful. The patient became quite anxious for a moment and then
went completely "blank." He came in the next session with a bland denial
of the emotional relevance of what had developed in the session before,
and once again the same cycle started all over, with repetitive declarations
of his boredom and the ineffectiveness of analysis [p. 245].
Kernberg believes that the process illustrated in this case is a reflec

tion of the narcissistic patient's inability to depend on the analyst.
Kernberg's continual interpretations of the patient's need to devalue
the analytic process were geared to making conscious the patient's
fear of acknowledging his need for the analyst because of his rage and
envy of him. As soon as the patient would begin to feel he had
received something useful from Kernberg, he became anxious and
then dissipated the anxiety by devaluing Kernberg and his offerings.
Otto Kernberg 233
It is the analyst's task to persist in the interpretation of these seem

ingly endless complaints and not be driven by them into collusion
with the patient by concluding that the patient is untreatable. While
many analysts believe that a patient who remains superficial and
seemingly uninvolved in treatment for a prolonged period of time is
unable to develop a transference and is untreatable, Kernberg's view
is that the narcissistic patient is in fact expressing an intense transfer
ence based on "devaluation, depreciation, and spoiling." Conse
quently, the interpretation of the negative transference is even
more crucial in the treatment of the narcissistic personality than in the
treatment of other borderline patients.
The most extreme form of narcissistic resistance is found in patients
who suffer from "malignant narcissism" (Kernberg, 1984). As dis
cussed earlier, the psychological organization of such patients tends to
be consumed by efforts to defeat others; hence they make triumph
over the analyst's efforts their primary treatment goal. Their sadism is
so gratified by such a "victory" that they are willing to lie and may
even become dangerously destructive and self-destructive in order to
achieve such a "triumph." Such patients self-mutilate and even make
serious suicide attempts in order to feel the exhilaration of defeating
the analyst. The first technical priority with such a patient is to con
front and limit both destructive and self-destructive behavior; other
wise, the treatment cannot proceed. The second priority is to confront
and interpret, if possible, the patient's lying, for this too renders treat
ment impossible. Kernberg acknowledges that once a firm stand is
taken against lying, the patient will frequently regress to a paranoid
transference. Nonetheless, in Kernberg's view, the analyst must be
able to withstand and work through this development because the
alternative is unworkable.
If these two threats to the treatment can be controlled, the treatment
of malignant narcissism proceeds to the major dynamic issues: the
patient's envy of and inability to depend on the analyst. These are typ
ical transference issues for most narcissistic patients, but in cases of
malignant narcissism they have a special quality: such patients have a
need to destroy the analyst psychologically. They attempt to defeat
the analyst's interpretations and to rob him of his most valued posses
sions, whether intellectual or physical. They set out to defeat the ana
lyst whenever the treatment threatens to progress. When the analytic
work finally seems to be going well after endless complaints of the
lack of progress, such a patient will suddenly dismiss it all with an
attitude of triumph. Kernberg attributes the intensity of this desire
to defeat the analyst to the patient's envy of the analyst's nurturing
Chapter 5
qualities and to his or her freedom from the same pathology that
enslaves the patient. The analyst's task is to interpret both the uncon
scious need to defeat the analyst and the envy that motivates it. It is
also crucial that the analyst not submit to the grandiose self but
instead take a firm stand on reality by interpreting the patient's distor
tions, even though the patient will become enraged and may act out in
ways that threaten to terminate the treatment. Kernberg does not
agree with Winnicott (1954b) and his followers (for example, Khan,
1960), who, as we saw in chapter 4, advocate tolerance of paranoid
regressive episodes. In Kernberg's view, the analyst's firm stance and
persistent interpretation of transference distortions reestablishes the
analytic framework and eventually allows patients to acknowledge
the unreasonable nature of their aggressiveness, resulting in guilt and
concern for others. At that point the establishment of more normal
object relations and superego structures begins to take place.
Because the patient treats the analyst as an extension of the self,
consideration of the countertransference is even more crucial in the
treatment of the typical narcissistic patient than in that of most bor
derline patients. In Kernberg's view, when a narcissistic patient con
sistently devalues the treatment process the analyst's emotional
response is a good indicator of the patient's hidden intention. In the
clinical vignette described earlier, Kernberg's feeling of being deval
ued as useless and silly was the clue to his interpretation that the
patient's intention was to make him feel worthless. The most difficult
and typical countertransference is the analyst's feeling of being con
trolled and devalued. The danger, in Kernberg's view, is in analysts
acting out their countertransference anger by rejecting the patient in
retaliation for the rejection they experience. Although analysts cannot
always assume that everything they experience is a reflection of the
patient's current issues, Kernberg does point out that because narcis
sistic patients treat the analyst as an extension of themselves, the
countertransference does tend to be much closer to the patient's affec
tive life than is normally the case. While other sources of data must
always be used, the countertransference tends to be a more significant
indicator with narcissistic personalities than with most other patients.
Confrontation with the rage, envy, and fear of dependence under
lying the grandiose self, if pursued persistently, eventually leads to
the eruption of these highly anxiety-provoking affects (Kernberg,
1975). The result is an upsurge of hatred and its projection, leading to
an intense negative transference. The pathological internalized object
relationships that the grandiose self split off from awareness now
become activated in the transference. The analyst becomes the sadistic
Otto Kernberg 235
parental object, but to defend against this awareness the patient re-
idealizes the analyst. In this phase of treatment of the narcissistic per
sonality, the patient's identification will rapidly alternate among the
idealized self- and object images and the negative self- and object
images while the analyst is perceived as the complementary represen
tation (Kernberg, 1984). For example, the patient may identify with
the infantile, victimized self while the analyst is experienced as the
sadistic mother, or the patient may assume the role of the idealized
father and see the analyst as an empty and greedy child, longing for
him. These split-off object relations units had been hidden under the
grandiose self and now appear in response to its interpretation. In
Kernberg's view, this development can take as long as three years to
appear and should be seen by the fifth year if the grandiose self is ana
lytically resolvable. In this advanced stage of the psychoanalysis of the
narcissistic personality, the treatment appears to be much like that of
other borderline patients. The focus is on the interpretation of the
quickly shifting, intense split object relations units.
Eventually, as the negative transference and oral rage are worked
through, patients are forced to become aware that the hated, feared
mother/analyst is a projection of their oral rage and is the same as the
idealized, longed-for mother/analyst whom they wish to be rescued by
(Kernberg, 1975). Kernberg agrees with Klein (1937) that the treatment
of splitting involves the resolution of the depression resulting from the
awareness that the hated object is the loved object. This realization ush
ers in a crucial period of treatment. Patients now feel guilty and
despairing for having hated the analyst/mother whom they also love
and long for. The guilt results in depression and may be strong enough
to elicit suicidal ideation. It is crucial that analysts be alert to the con
text and reasons for the depression. If they are not, the depression may
be mistakenly viewed as regression, rather than progression. Kernberg
sees depression as not only a positive movement in therapy but also
a significant structural shift, as the patient is able to experience true
guilt for the first time and mourning can now take place. It is critical
that the analyst understand that new psychological structure is being
formed and interpret the patient's depression as guilt over having
injured, even if only in fantasy, the object of love and dependence.
That the patient can now experience this depression reflects three
crucial therapeutic transformations. First, the ability to experience
ambivalence indicates movement to the level of whole object integra
tion. Because whole-object integration is the necessary condition for
the formation of the structured tripartite ego, the resolution of guilt
and ambivalence results in the development of the ego-superego-id
Chapter 5
structure of the psyche. Second, the analyst is experienced as a sepa

rate, independent person for the first time, allowing the patient to
show genuine interest in others as separate, "real" people. Third, and
perhaps most important, the resolution of the narcissistic resistances
results in normal infantile narcissism and now dependence on the
analyst. Interpretations can now be used for understanding, as
opposed to being "stolen" greedily as magical "food" or rejected to
defend against envy. A benign cycle takes place in which the utiliza
tion of interpretations leads to an internal richness that replaces
emptiness and gradually alleviates envy.
The best description Kernberg (1975) provides of the treatment
process with the narcissistic patient is the analysis of an architect in
his late thirties. For the first three years of analysis the patient showed
the typical narcissistic paradigm. He initially idealized the analyst and
then shifted to a continual oscillation between grandiosity and ideal
ization. The interpretation of these narcissistic defenses led initially to
the revelation of intense envy and competitiveness and later to oscilla
tion between "oral demandingness and anger to longing for a depen
dency on a loving, protective father-mother image, and strong guilt
feelings for his attacks on the analyst" (p. 304).
This transference paradigm shifted to a more stable dependence
on a loving father image, and it was only at this time, after three
years of analysis, that the patient came to depend on the analyst as
a real person. He now began to miss him on weekends and separa
tions and for the first time felt depression and mourning in his
relationship with the analyst. The patient reacted to this phase by
withdrawal and feelings of emptiness, and the sessions appeared to
revert to the previous stage of narcissistic resistances. However,
the material was replete with references to the patient's sadistic,
withholding mother, and he became aware that he was identified
with her while he treated the analyst as his infantile self. The inter
pretation of this new transference paradigm brought about a fur
ther deepening of the realistic dependence on the analyst. Kernberg
described the ensuing process this way:
The patient now saw [the analyst] as a protective, loving father toward
whom he could turn for the gratification of his dependent childhood
needs; and he now felt he could abandon himself to the analytic situation.
. . . Now, for the first time, the patient became aware of how his entire
attitude toward the analyst had been influenced by his basic conviction
that no real relationship would ever occur between him and the psy
choanalyst. . . . A year later the full development of oedipal conflicts
Otto Kernberg 237
emerged in the transference, and the analysis acquired features of the

usual resistances and manifestations of these conflicts [p. 3051.
This vignette shows how a prolonged period of narcissistic resis

tance was eventually "undone" by persistent interpretation of the
grandiose self and idealization defenses. The result was an eruption
of oral rage and envy that was worked through by interpretation,
and followed by the crucial stage of object integration to which the
patient responded by narcissistic withdrawal. Even though this stage
appeared to erase the treatment gains, it was a temporary regression
stimulated by the anxiety of ambivalence toward the analyst/mother
whole object. Depression and mourning were signs of forward move
ment. The fact that interpretation quickly moved the process forward
indicates that progress had taken place. At that point the dependence
on the analyst took root allowing the patient to experience his own
emotional needs as well as those of his family. This case illustrates
Kernberg's view that when the patient is able to depend on the ana
lyst, the narcissistic defenses have been resolved and the analysis
moves to a neurotic level.
By way of summary, it can be seen that Kernberg's view of the
treatment of the narcissistic personality eventually becomes similar to
the treatment of other borderline patients. The most striking differ
ence is the prolonged period of time devoted to the interpreting and
working through of the narcissistic resistances that precedes the reve
lation of the borderline dynamics of pathological internalized object
relationships. The other major difference is the prominence of the
envy that emerges in the pathological constellation. This envy leads to
especially difficult resistances and regressive swings as the patient is
unable to depend on the analyst and is therefore resistant to accepting
interpretations. Because both these factors make the defenses of nar
cissistic patients especially difficult to resolve, Kernberg prefers psy
choanalysis proper for these patients whenever possible to give the
best possible chance for the resolution of the narcissistic resistances.
SUMMARY AND CRITIQUE
Kernberg has created the most comprehensive object relations theory

of development, psychopathology, and treatment yet developed. His
approach is a systematic application of both object relations and ego
psychology developmental theory to pathology and treatment. He has
been able to enhance the value of the Kleinian object relations model
and its understanding of defense mechanisms without falling into the
238 Chapter 5
most serious errors of the Kleinian school (see chapter 3). No other
theorist has succeeded so well in synthesizing the work of Jacobson,
Mahler, and Klein and in integrating object relations theory with the
structural concepts and technical principles of ego psychology.
Further, Kernberg has developed an object relations model of treat
ment consistent with the technical requirements of traditional psycho
analytic therapy. In this respect, his thought stands in opposition to
the work of Fairbairn, Guntrip, and Winnicott, all of whom ultimately
recommended modification of the classical analytic stance in the inter
ests of becoming more of a real person to the patient. At this point, a
critical assessment of Kernberg's work is in order. Three categories of
difficulty with Kernberg's views will be discussed.
First, there are conceptual problems with some of Kernberg's dis
tinctions between psychopathological conditions and their clinical
implications. Foremost among these difficulties is confusion sur
rounding his conceptualization of a subgroup of narcissistic patients
who function on the overt borderline level. By Kernberg's definition,
narcissistic personalities are borderline patients who use a grandiose-
self defense to conceal the underlying personality organization. The
conceptualization of a subgroup for which this is not true contradicts
Kernberg's own definition of narcissistic personality organization.
Kernberg contends that patients in this subgroup, despite overt bor
derline functioning, are narcissistic personalities because of the pre
dominance of envy and the presence of a narcissistic personality
structure, although its functioning is ineffective. However, these crite
ria change Kernberg's definition of the narcissistic personality organi
zation and make this subgroup virtually impossible to distinguish
from other borderline patients who use omnipotence as a defense.
Further, since psychoanalysis is not recommended for this group—
sometimes not even expressive psychotherapy—its inclusion in the
category of narcissistic patients is even more dubious.
Perhaps more significantly, there is a problem with Kernberg's rec
ommendation of psychoanalysis proper for narcissistic patients and
expressive psychotherapy for other borderline patients. The latter are
deemed too weak in ego strength and boundaries to withstand psy
choanalysis, yet psychoanalysis is recommended for the narcissistic
patient, who is said to have an underlying borderline personality
organization. If narcissistic patients are really borderline personalities,
their weak ego structure should contraindicate psychoanalysis for
them as for other borderline patients. Kernberg's contention is that
when the grandiose-self defense is worked through in treatment, the
underlying borderline personality organization becomes manifest.
Otto Kernberg 239
According to Kernberg's conceptualization of treatment, expressive

psychotherapy should be the treatment of choice at that point.
Kernberg's rationale for recommending psychoanalysis for the nar
cissistic patient is that only the most intensive form of treatment can
hope to dissolve narcissistic defenses. While this may be so, it does
not address the contradiction inherent in the assertion that psycho
analysis can be effective in the treatment of the now-apparent ego
weaknesses of the narcissistic patient's underlying borderline per
sonality organization whereas it is not considered effective with
other borderline patients.
These considerations lead to a general problem with Kernberg's
concept of applying different treatments to different of levels of per
sonality organization. The neurotic personality organization is consid
ered appropriate for psychoanalysis proper, unlike severe character
pathology, which requires the integration of split object relations.
However, as we have seen, Kernberg views psychoanalysis as a
process of breaking down the personality into its component object
relations units, and then reintegrating them into a new structure.
Given this concept of the analytic process, the difference between the
treatment of neurosis and severe character pathology lies primarily in
the setting of limits. Yet, as we saw, many borderline personalities do
not require the setting of limits. Treatment approaches that Kernberg
presents as clearly different are more alike than he believes. The simi
larity of treatment approaches is underscored by the neurotic patient
discussed earlier who showed no more simultaneous awareness of
different object relations units than do most borderline patients. The
analytic process for the neurotic patient involved undergoing different
phases in the enactment of split-off object relations units, a process
that is difficult to distinguish from the treatment of the borderline
patient who does not act out severely. The common occurrence of this
process in the treatment of neurotics led Grotstein (1986) to conclude
that splitting is ubiquitous rather than a particular mechanism of
severe psychopathology (see chapter 3). Kernberg presents his cate
gories of patients and treatments as discrete groupings, whereas the
differences seem to be of degree and not kind, and the criteria for
making the distinctions are much less clear than Kernberg contends.
The second questionable aspect of Kernberg's theory is his adher
ence to the concept of drive. In Kernberg's view, a drive can only be
expressed through an object relationship. The embeddedness of drives
in object relations raises the question of why the concept of drive
should be retained at all. The primary reason Kernberg gives for
maintaining the drive concept is that a pure object relations model
240 Chapter 5
denies the importance of the aggressive drive, which seeks not to

obtain an object but to eliminate it. But, is aggressiveness a drive? As
mentioned earlier, Kernberg distinguishes between an instinctive
response as an inborn, discrete self-preservative behavior, and a drive.
By definition, a drive is a cyclical biological impulse that motivates
behavior to achieve gratification. Kernberg uses the concept in this
sense when he contends that the aggressively acting-out patient is
receiving "instinctual gratification." As pointed out in the critique of
Klein (see chapter 3), sex, warmth, hunger, and thirst all fit this model,
but aggression does not have such a biological rhythm (Scott, 1958).
Aggression does fit the model of an instinctive response, as it is an
innate disposition that is evoked when circumstances warrant. How
ever, to conclude from this that it is a drive seeking gratification is to
disregard Kernberg's own distinction between instinctive response
and drive. The evidence that aggression does not operate as a drive
appears to eliminate Kernberg's only justification for differentiating
his theory from a pure object relations model.
Similarly, Kernberg treatment strategy does not justify postulating
an aggressive drive. Kernberg applies this concept to both the setting
of treatment parameters and the interpretive process. One need not,
however, assume that aggression is "instinctually gratifying" in order
to justify limiting destructive behavior. It is sufficient that the behav
ior is harmful. Kernberg (1975) provides two clinical illustrations of
limiting "instinctual gratification." One case was the patient men
tioned earlier who screamed at her therapist interminably although
she was calm and relaxed outside the sessions. The therapist limited
her behavior, thus creating more conflict outside the sessions and an
increase of anxiety. One can understand the material as easily from a
purely object relations point of view: the patient may have been enact
ing a split object relationship from her past with reversed roles so that
she was the abusive parental figure and the therapist was placed in
the role of the victimized child/patient. Having achieved mastery
over the childhood abuser/therapist, she was then able to relate nor
mally to other people. When the therapist prevented the enactment of
this relationship, the sense of mastery over trauma was taken away
and the need to enact it infiltrated other relationships. While one can
not make a definitive formulation from the material presented in
Kernberg's vignette, it is sufficient for the present purpose to note that
it can be understood without the assumption of drive gratification.
In another case Kernberg presents as an illustration of limiting
"instinctual gratification," the patient angrily demanded an increase
in hours. When interpretation was to no avail, the therapist refused to
Otto Kernberg 241
increase the hours and made a change in the patient's behavior a con
dition for the continuation of the treatment. The patient's behavior
changed markedly in a few days, and he admitted that he had enjoyed
his expression of anger. The positive response to limit setting contra
dicts a drive interpretation of the patient's behavior. According to the
postulates of psychoanalysis, if the expression of a drive is blocked,
frustration will be experienced and the drive derivative will seek
expression in an indirect, pathological form. Apparently this did not
happen in this case. Furthermore, the satisfaction experienced by the
aggressive expression does not imply "instinctual gratification":
enactment of an object relationship can be enjoyed for any number of
reasons, depending on its meaning to the patient.
Kernberg applies the concept of an aggressive drive to the interpre
tive process in his recommendation that the therapist focus on the
negative transference. We have seen that he attributes the abusive,
hostile acting out of the borderline patient toward the therapist to
"excessive aggressiveness." However, in Kernberg's view, as drives
are expressed only through object relationships, the negative transfer
ence is the enactment of a split object relationship, and the therapist's
task is to discover and label it. The conceptualization of the aggressive
object relationship as a drive does not add to the interpretation and
could potentially misdirect it. To give but one example: when Miss A
devalued her therapist as "provincial," intellectually inadequate,
unsophisticated, and lacking in self-assurance, she was enacting her
mother's superior, devaluing attitude toward her as a child and caus
ing the therapist to feel the sense of inadequacy and despair she had
felt. Kernberg interpreted this transference paradigm as an enactment
of her internalized mother-child object relationship. To equate this
transference enactment with an aggressive drive does not contribute
to the interpretation, and to interpret the patient's hostility as the act
ing out of excessive aggressiveness would be to miss its object rela
tions significance. To Kernberg's credit, he does not suggest the latter
but proposes the object relations interpretation. However, having
done so, he obviates the need to invoke aggression as a drive.
A related problem with Kernberg's concept of the aggressive drive
is his failure to distinguish between aggression and hate. When
Kernberg develops his view that the earliest object relationships are
divided into the libidinal and aggressive object relations units, he
identifies libido with love and aggression with hate. However, aggres
sion is an energetic expression that can occur in a variety of affective
contexts such as joy, excitement, mastery, assertiveness, or hate. The
degree to which aggression is manifested as hate is often the degree to
242 Chapter 5
which it tends toward the pathological. As we will see in chapter 6,

Kohut (1977) makes the distinction between the healthy assertive
ness of the normal child and the pathological rage of the narcissistic
personality disorder. The baby in distress is clearly in pain, and
Kernberg is quite justified in presuming that such states of "unplea
sure" or pain are in marked contrast to the infant's pleasurable
states. However, it is an unjustified inference to conclude that the
infant feels "hatred" whenever it feels pain. Hatred appears later,
when the child is able to target an object of its frustration. Aggres
sive acting out of hateful feelings is one type of expression of
aggression, the most negative form it can take.
To identify the intense hatred and destructive expressions of the
borderline patient with childhood aggression is to equate pathology
with normal childhood, a mistake Kernberg accuses Kohut of making
with the latter's concepts of grandiosity and idealization. While
Kernberg (1975) criticizes Kohut for failing to distinguish normal and
pathological idealization, Kernberg appears to have made a similar
error in equating hatred and aggression. The abusive, destructive hos
tility of the borderline personality organization is not an excess of
joyful aggressive assertiveness; it is a different psychological phenom
enon. Thus, excess aggression of people who are assertive in the
achievement of life and work goals is qualitatively different from the
hostility of the borderline patient. The aggression of the borderline
patient is pathological not because of its quantity but because of its
hateful, destructive quality. To view hateful, destructive aggressive
ness as somehow primary, and healthy assertiveness as a sublimation
of the hateful state is to pathologize human motivation in ways that
are not consonant with recent developmental research into infancy
and childhood (Lichtenburg, 1983,1989).
Stern's (1985) compilation of the evidence from infant research
leads to the conclusion that the notions of hatred and "bad" are sym
bolic forms beyond the ken of the infant. Stern points out that the
infant has four to six feedings each day with varying degrees of plea
sure and that the infant's cognitive ability enables it to discern face
and breast across these varied tones of pleasure more easily than it is
able to organize experiences into pleasure and unpleasure and then
conceptualize these states as "good" and "bad." It is perhaps unneces
sary to state that hatred requires still one more inferential step. The
evidence from infant research indicates that "good," "bad," and
"hatred" are later conceptualizations that are possible only after
speech has developed. Stern does suggest however, that infant
research allows one to infer that infants experience pleasure and
Otto Kernberg 243
unpleasure in the early phase and that these experiences will be clus
tered around their "hedonic tone." However, he points out that the
infant has many such "working models of mother" and that there is
no basis for giving primacy to categorization according to hedonic
tone. Stern concludes that splitting is not a normal state of infancy but
a later construction that lends itself to pathological conditions.
Given that aggression does not evince the biological rhythm of
drives, that drives in Kernberg's view "find expression" only through
object relationships, and, finally, that the presumption of aggression
as a drive seems clinically unnecessary and confusing, Kernberg's
object relations model does not require the retention of a drive con
cept. It does not fit Kernberg's metapsychology and theory of person
ality development, and it renders his clinical theory needlessly
confusing. In fact, the drive concept simply risks interfering with the
consistent application of his object relations model of treatment. The
only purported application of the concept is to explain limit setting,
which, we have seen, can be explained without a drive concept. Even
at the oedipal level, Kernberg's approach is to analyze the constituent
object relations that compose the psychological organization. It is the
apparent irrelevance of the drive concept to the object relations view
of psychoanalytic treatment that led Greenberg and Mitchell to con
clude that Kernberg was being "political" in his retention of the drives
(Greenberg and Mitchell, 1983).
This discussion of the drives leads to a third cluster of questions
regarding Kernberg's treatment principles for pathological aggres
sion. His primary postulate of confronting the patient with split-off
aggression to "undo" the splitting process is not sustained by his
own clinical illustrations. In the clinical vignettes he offers to demon
strate this technique, the patient frequently regresses. In one case
illustration of "undoing" splitting discussed earlier, the therapist
confronted the patient with her alcoholic binges and the patient
needed to be rehospitalized. The therapist felt he was dealing with a
"real person" for the first time—and may well have been. However,
given the patient's reaction, this feeling on the part of the therapist
does not demonstrate the effectiveness of the technique; follow-up
data confirming the expected long-term beneficial results are neces
sary. While Kernberg is convinced of the importance of "undoing"
splitting by confrontation in such cases, the clinical vignettes he
reports do not justify his certainty. The fact that the patients often
become more symptomatic may well be a product of ill-timed or
misconceived interpretation. For example, if the aforementioned
patient was drinking for a reason other than anger at the therapist,
244 Chapter 5
she may well have regressed in response to feeling misunderstood.

To support his claim for the effectiveness of his method, Kernberg
often refers to the findings of the Menninger Research Project on
Psychotherapy and Psychoanalysis (Kernberg et al., 1972). This
impressive research project is one of the few well-controlled studies of
psychotherapy with severely disturbed patients, and its findings pro
vide no evidence relevant to the validity of Kernberg's approach to
excessive aggression or aggressively fueled defense mechanisms.
More significantly, the study is not as supportive of Kernberg's claim
that expressive psychotherapy is the treatment of choice for borderline
patients as he seems to imply. The Menninger study only compared
supportive psychotherapy to psychoanalysis proper; there was no
group receiving expressive psychotherapy. The basic findings were
that patients with low initial ego strength (borderlines) did poorly in
both treatment modalities and patients with high initial ego strength
did well in both modalities. In both groups, patients receiving psycho
analysis did better than patients receiving supportive psychotherapy.
However, there were some patients in the latter group who received
"supportive-expressive" psychotherapy, and the low initial ego
strength group did better with this modality than either of the others.
Kernberg deduces from this finding that these poorly functioning
patients require a special type of treatment: modified psychoanalysis.
It needs to be underscored that no group in the study was given this
type of treatment. Kernberg's conclusion is an inference from the data,
not an outgrowth of it. While he recognized that he was making such
an inference, he pointed to his clinical data to support his claim for
modified psychoanalysis (Kernberg et al., 1972). But, as noted, his clin
ical data are wanting. It must be concluded that despite the impres
sive, consistent system Kernberg advocates for the treatment of the
borderline patient, the clinical and research data he offers in support
of it do not justify such confidence.
While Kernberg's approach to the undoing of splitting and the han
dling of the negative transference is not convincing, he provides better
evidence for his overall strategy of labeling split object relationships
and integrating them via interpretation. For example, when Miss L's
two primary transference paradigms—the wish to be raped and killed
and the infantile dependence—were brought together, she began her
improvement. This integration was achieved by interpretation rather
than an undoing by confrontation, and the mutative effects of the
interpretive process are impressive. Three points are relevant in the
present context. First, the pathology in this case lay not in excessive
aggressiveness but in the splitting of two conflictual object relations
245
units. This fact would seem to belie Kernberg's claim that splitting
is always a product of excessive oral aggressiveness and that the lat
ter is the root of pathology in the borderline personality organiza
tion. Second, the splitting of Miss L's object relations units cannot
be conceptualized as a split between good and bad. Both object rela
tionships had positive and negative qualities, and both were fraught
with conflict. In fact, very few of Kernberg's clinical illustrations of
splitting can be readily formulated as good and bad. Miss L repre
sents the more typical borderline patient who splits between two
highly conflicted object relationships, each of which is threatened
by the other. Third, this case, like others without acting out of
excess aggressiveness, was managed by interpretation alone and
showed great improvement.
The same principles apply to Miss N, the patient who saw Kern
berg as either a harsh, dangerous masculine figure or a soft, slightly
depressed feminine figure. Despite Kernberg's attempts to fit these
transference paradigms into his "all good" and "all bad" scheme, both
object relations units had good and bad qualities. They were split
apart, but not organized by good and bad. In this case, like the case
of Miss L, there was no indication of "excessive aggressiveness" and
no discussion of acting-out a negative transference. In cases of this
type—in which object relations units are split, but not into good and
bad—Kernberg seems to apply his interpretive principles of object
relationship enactment most clearly and consistently, and one can
most easily see therapeutic benefit. There are some patients, such as
Miss B, whose splitting is organized around good and bad, but they
appear to be but one variant of split ego organization.
In summary, Kernberg's clinical illustrations fit his concept of
object relationships as the root of the borderline personality organiza
tion but not his equation of splitting with good and bad. In the cases
discussed, one can see that the pathological splitting process involves
complex object relations units that are split apart for a variety of rea
sons. It would be a major oversimplification to reduce the pathology
to excess aggressiveness. The splitting process in both Miss N and
Miss A was not motivated by excess aggressiveness but by the need to
keep apart perceptions of self and object that were too confusing and
difficult to integrate.
Finally, it should be noted that this critique of the drive component
of Kernberg's thought is not meant to suggest that his theory of the
analytic process is implicitly social or interpersonal. The current
school of interpersonal psychoanalysis, discussed in chapter 7 and
represented by such theorists as Gill and Mitchell, conceptualizes the
246 Chapter 5
analytic process as interpersonal in the sense that patient and thera

pist are engaged in a mutual relationship in which each party continu
ally e nacts and reenacts roles (Gill, 1983; Greenberg and Mitchell,
1983; Mitchell, 1988). For Kernberg, role enactments are a playing out
of the patient's drama. The therapist's behavior will often fit a pre
scribed role, but the interaction follows from a script written by the
patient's unconscious. Kernberg has managed to create a conceptual
ization of psychoanalytic treatment that takes into full consideration
the enactment of role relations between patient and therapist without
becoming an interpersonal alternative to the psychoanalytic process of
an analyst analyzing a patient.
CHAPTER 6
The Work of Heinz Kohut
H e in z K ohut b e g a n h is m a jo r t h e o r e t ic a l in n o v a t io n s in p s y c h o -
analysis with the effort to fill what he, much like Kernberg, felt was a
void in the psychoanalytic theory of psychopathology between the
structural neuroses and the psychoses. Initially, he applied his self-
psychological theory to the narcissistic disorders, which he felt were
the primary syndromes lying between these two classes of pathology,
and accepted the drive-ego model for the neuroses. This concept of
self psychology became known as self psychology "in the narrow
sense." Eventually, Kohut expanded his theoretical viewpoint to self
psychology "in the broad sense," which included his reconceptualiza
tion of the neuroses as well as the therapeutic action of psychoanaly
sis. This shift entailed considerable conceptual change that Kohut did
not always acknowledge. For this reason his work can be confusing
and lends itself to different interpretations. While Kohut never altered
his initial views of narcissistic pathology, he broadened his theoretical
constructs to envelop all of psychoanalytic theory. His new theory,
self psychology, became the second major school within the frame
work of psychoanalysis. Adherents of self psychology differ as to
whether their model is an object relations theory. For example, Wolf
(1988) views self psychology as an intrapsychic theory and eschews its
categorization as object relations, whereas Bacal and Newman (1990)
conceive of self psychology as a variant of object relations theory. Self
psychology is classified as an object relations theory here because it
views development and psychopathology as rooted in the internaliza
tion of objects and thus fits the definition of object relations theory as
presented in chapter 1.
Kohut (1959, 1982) viewed psychoanalysis as the science of com
plex mental states. He pointed out that any science is defined by the
method it uses to gather its data. Psychoanalysis, as the psychology of
complex mental states, uses empathy and introspection. Kohut
defined empathy as vicarious introspection, the means by which we
understand others' communications. It is our ability to adopt the
247
248 Chapter 6
stance of others from their viewpoint that allows us to understand

them. The knowledge we gain by looking at the world through the
lens of others constitutes psychological knowledge. Whereas the nat
ural sciences gather their data by observation, acquiring knowledge of
the physical world through the senses, psychological knowledge is
won by empathy and introspection. We understand other people not
by observing their behavior but by empathizing with their psychologi
cal states. Knowledge gathered by extrospection is not psychoanalytic
data and cannot inform psychoanalysis as science or therapeutic pro
cedure. From a theoretical point of view, Kohut hoped to delimit the
type of concept admissible in psychoanalytic discourse. He advocated
the elimination of concepts such as dependence from psychoanalytic
consideration because he felt that such concepts could be either bio
logical or sociological but never psychological (Kohut, 1982). By defin
ing psychoanalysis as the science of empathy and introspection he
hoped to found it as a pure psychology, independent of biology or
any other discipline. The implication for treatment is that free associa
tion is replaced by empathic immersion in the patient's psychological
world as the primary investigative tool (Kohut, 1959,1977, p. 303).
In accordance with his emphasis on empathy and introspection,
Kohut (1984) distinguished "experience near" from "experience dis
tant" theory: theory based on the empathic grasp of what patients
experience at any particular time is experience near whereas theory
that uses an external frame of reference is experience distant. That
is, considering certain behaviors normal or abnormal is an experi
ence-distant approach, whereas grasping empathically the meaning
of the behavior is an experience-near approach. The central point of
Kohut's therapeutic prescription is that the analyst must always be
attuned to the experience near, the meaning of the behavior, even
though his empathic understanding has to be informed by experi
ence-distant guidelines of normality. Kohut believed that classical
psychoanalytic theory had become too removed from patients'
experience and that psychoanalytic theory must return to its roots
in patients' subjective experience.
Because Kohut's views shifted dramatically, and were continually
evolving, even immediately prior to his death in 1981, his work is best
approached historically. Consequently, after a brief account of his
developmental views, which provides the conceptual framework for
his general theory of psychopathology, his formulations of narcissistic
pathology will be discussed in detail. We will then examine their
expansion into self psychology as a global theory that now forms the
basis of a separate psychoanalytic school.
Heinz Kohut 249
DEVELOPMENT
Kohut (1966, 1971) accepted Freud's (1914) theory that the initial
stages of infancy are autoerotism and primary narcissism, which is
defined by the libidinal investment of the self, and that the infant
emerges from the latter stage when it recognizes that its source of sup
plies comes from outside itself. At this point, says Kohut, following
Freud, much of the narcissism from the stage of primary narcissism is
now invested in the parental figure, resulting in the idealization of the
object. In the resolution of the oedipal complex, the parental figure is
internalized to form the superego-ego ideal structure. Thus, the per
sonality structure is a result of the reinternalization of primary narcis
sism after a "deflection" through the idealized parental object. Up to
this point, Kohut follows Freud.
According to Kohut (1977), the earliest phase of infancy is a prepsy-
chological stage characterized by physiological needs and tension
states without awareness. However, he believed that one could justifi
ably view even this early prepsychological state as a self in statu
nascendi because the infant is treated as though it were a self. The
empathic maternal environment responds to the baby as a virtual self,
and so begins the process of self formation. Even the first maternal
contact begins the process that will result in the birth of the self. Thus,
ministrations of the maternal environment constitute the first "selfob
ject" in the infant's experience. By this term Kohut refers to the object
as fulfilling a necessary (that is, psychologically life-sustaining) func
tion that the self is unable to perform for itself. (Initially Kohut
hyphenated this term, but since he eliminated the hyphen in all his
later work and his followers have continued this usage, the hyphen
will be eliminated here as well.)
The early maternal environment provides the child's first experi
ence of empathy; by means of empathy the mother of the neonate
knows the child's psychological states and responds to them.
Empathy, the tool by which we know others, is also the origin of
psychological life in Kohut's (1977) view: without the empathic
mother responding to the infant as though it already had a self, the
infant would not develop a sense of self. The child is born with
innate potentials, and the environment responds selectively to them,
thereby channeling innate givens into a "nuclear self." In short, the
birth of the self is a function of maternal empathy mobilizing the
child's constitutional endowment.
Psychological life begins with emergence from primary narcissism
and is stimulated by the inevitable failure of the parents to satisfy
Chapter 6
perfectly the child's narcissistic needs. According to Kohut (1971), the

child now becomes aware of its vulnerability and attempts to recap
ture the lost bliss of the narcissistic state in two ways. First, the child
effects a renewed libidinal investment in the self, which Kohut origi
nally called the "narcissistic self" and later changed to the "grandiose
self." This self has grandiose and exhibitionistic needs, but since it is
too weak and vulnerable to satisfy them itself, it requires an object for
confirmation. A crucial function of the parental object in this phase is
to "mirror" the vulnerable, grandiose self. The parent who mirrors the
child's need for admiration and approval is performing a necessary
function and thereby bolstering the child's sense of self. The child who
receives such mirroring does not view the parent as an autonomous
individual but as a means for strengthening its sense of self; this is
why Kohut terms such parental behavior a "selfobject" function.
Some of the narcissistic libido is invested in the idealization of the
parent, which results in what Kohut called the "idealized parental
imago." Attachment to this imago provides an additional source of
narcissistic gratification. The child oscillates between recognition of
the parent as "other" and resubmergence into a primitive merger
state. Out of this oscillation, the idealized parental imago is estab
lished. In supplying this narcissistic function the parent once more is
viewed not as an individual in his or her own right but as a means
by which the child can achieve an enhanced sense of well-being.
Availability for idealization, then, is the second selfobject function
performed by the parent in the first years of life. The grandiose self
and the idealized parental imago are the transformations of primary
narcissism that form the archaic self in the preoedipal phase. While
Kohut recognized that either parent may be involved in mirroring the
grandiose self or allowing idealization, he did believe that one could
roughly equate the mother with the mirroring function and the father
with the need to idealize.
Gradually, as further disappointments in the parent are recognized,
the idealized parental imago is given up. Every blemish on the ideal
ized parental imago leads to its internalization until the ego ideal is
firmly established within the psyche. The process by which this occurs
is an almost imperceptible yielding of the idealized parental imago
and its transformation into psychological structure, the ego ideal.
Kohut calls this process "transmuting internalization," a term he
applies to any internalization process that results in new psychologi
cal structure. In this case the idealized parental imago is transformed
into ideals that the child hopes to realize.
Kohut follows Freud in seeing superego structure as the result of
Heinz Kohut 251
oedipal loss. However, unlike Freud, he sees the idealization of the

superego as resulting not from its content but from its genesis in
primary narcissism. According to Kohut (1966):
That the original narcissism has passed through a cherished object

before its reinternalization and that the narcissistic investment itself has
been raised to the new developmental level of idealization account for
the unique emotional importance of our standards, values, and ideals
insofar as they are part of the superego [p. 434].
Thus, the internalization of the idealized parental imago results in

the formation of the ego ideal, this new drive-regulating structure also
imparts vitality to the superego, which becomes idealized, life in gen
eral now incorporates goals worth striving for. The end result of this
process is an internalized set of ideals, cherished because of their
connection to lost primary narcissism. These ideals form, in Kohut's
terminology, one "pole" of the self.
The distinction between the process of superego formation and the
ego ideal implies a difference between the development of narcissism
and the development of object libido. Postulating these two separate
lines of development was critical for Kohut; it allowed him to account
for discrepancies between narcissistic equilibrium and superego for
mation. Kohut disputed Freud's U-tube analogy of narcissism in
which the more narcissistic cathexes are "sent out" to objects the less
love is available for oneself. He pointed out that love for the object
does not decrease self-esteem but, rather, enhances it. In the present
context it is unnecessary to summarize his argument; it suffices to
note that, for Kohut, narcissistic libido and object libido underlie sepa
rate developmental processes. Narcissistic cathexes that are invested
in the idealized parental imago result in the drive-neutralizing capac
ity of the psyche and in the idealization of the superego whereas
object cathexes result in superego formation. The implication of this
distinction is that the superego can be formed without being ideal
ized; in such a case, superego structure is present but has little mean
ing to the personality. We will see the value of this distinction in
Kohut's extensive discussion of patients with narcissistic personality
disorders who have integrated but inadequately idealized superegos.
The archaic infantile grandiosity has a different fate. It too is gradu
ally relinquished as parental responses inevitably fail to satisfy the
child's grandiose needs. As the grandiose-exhibitionistic needs are
optimally frustrated, they become transformed into realistic ambi
tions, forming the other pole of the self structure. Ideals and ambitions
252
constitute the "bipolar self." In Kohut's view, ambitions push the indi
vidual but they are not loved as ideals are. Narcissistic libido becomes
transformed from infantile, archaic grandiosity to realistic ambitions;
these ambitions drive individuals to achieve their loved ideals.
Between ambitions and ideals lie the individual's innate talents and
skills. Ambitions and ideals must be sufficiently realistic so that they
are within reach of one's talents, and the latter must be adequate to
achieve ambitions and fulfill ideals. Consequently, Kohut conceives of
talent as a "tension arc" between the ambitions and ideals that makes
their achievement possible. In normal development the infantile
grandiose self and idealized parental imago are transformed into real
istic ambitions and loved ideals so that a cohesive, vital, harmonious
self is capable of achieving a fulfilling life by realizing its ideals and
ambitions with its available talents.
Kohut's views on the Oedipus complex changed decisively during
the course of his theoretical development. In his early work Kohut
(1971) accepted the classical psychoanalytic view of the Oedipus com
plex as the crucial developmental phase for the pathogenesis of the
structural neuroses. At that point, he confined his contribution to the
delineation of narcissism between the phase of primary narcissism
and the oedipal period. His contention was that once the narcissistic
phase was overcome by the relinquishment of archaic selfobjects and
the emergence of an independent self, the child entered the oedipal
phase with the typical sexual and aggressive conflicts described by
classical psychoanalytic theory.
In his second book Kohut (1977) showed indications of shifting
away from strict adherence to the classical position, but he did not
delineate a clear alternative view. In subsequent publications, most
notably in his third book, it became clear that Kohut (1984) no longer
viewed the oedipal phase as a qualitatively different developmental
period characterized by drive, as opposed to narcissistic, conflicts. In
this revised developmental scheme, once the self is firm, vital, and
harmonious, it is ready to address the issues of the oedipal phase but
does not necessarily engage an Oedipus complex. The oedipal phase
involves phase-appropriate affectionate and assertive feelings toward
the parents; the Oedipus complex is the pathological outcome of unre
solved conflicts in the oedipal phase. The phase is universal, the com
plex is not. The child who has formed a strong sense of self via the
internalization of the preoedipal selfobjects, according to Kohut, joy
fully enters the oedipal phase. If the parental objects are empathic
with the affectionate feelings for the opposite-sex parent and the
assertive feelings toward the same-sex parent and if they respond
Heinz Kohut 253
with pride and affection of their own, the child will be able to inte
grate affectionate and assertive strivings into its self structure. The
integration of these affects enhances the ability of the self to achieve its
ambitions and ideals.
Departing from classical theory, Kohut (1984) ultimately contended
that sexual and hostile feelings in this phase are neither ubiquitous
nor healthy but, instead, a product of pathogenic parental responses.
According to Kohut, the drives are not inborn; disposition to affection
and assertiveness is inborn, but these feelings only become trans
formed into lust or hostility under pathogenic conditions. The normal
outcome of the oedipal phase, in Kohut's view, is not the acquisition
of mastery over the drives, but the ability of the self to achieve its
ambitions and fulfill its ideals. The key variable is neither the strength
of the drives nor their prohibition but the empathy of the selfobject
milieu.
The place of the oedipal phase in Kohut's final theory of develop
ment is very different from the decisive role it plays in classical
psychoanalytic theory for the ultimate health of the personality. In
Kohut's scheme previous phases are more crucial because if the self is
strong, vital, and harmonious on entering the oedipal phase, the fail
ure of the oedipal selfobjects can only impede the self's achievement
of its goals; it cannot weaken self structure.
It is clear that drives do not play a major role in normal develop
ment in Kohut's scheme. The crucial variable is the development of
the self, which is a product of the responses of the selfobjects of
childhood to the narcissistic needs of the child. If the self is cohesive
and vital there are no isolated lustful longings or hostile wishes. In
the healthy self, according to Kohut, both affectionate and assertive
feelings are experienced joyfully. To the extent that the self is defec
tive, affection becomes split off into lust, and self-assertion "breaks
down" into hostility. In his view, sexual and hostile wishes are not
inborn drives that must be tamed to achieve psychological organiza
tion; rather, the dispositions toward affection and assertion are
inborn, and hostility and lust are pathological products of their dis
tortion. The fate of affection and assertion is determined by the
strength of the self which, as we have seen, is a function of the
responses of the crucial selfobjects. If the selfobject responses are
empathic, phase- appropriate, and optimally frustrating, sexual feel
ings are the healthy outgrowth of affectionate feelings and aggres
sion is the joyful expression of self-assertion. If the responses of the
selfobjects are not empathic but are excessively frustrating or stimu
lating, either in the preoedipal or oedipal phases, self-development
254 Chapter 6
is arrested and pathology ensues. We now turn to the variety of

pathological outcomes of faulty selfobject responses.
PSYCHOPATHOLOGY
Kohut's views on psychopathology grew out of his developmental

theory. Since the crucial aspect of emotional development is the self
and since its well-being depends on selfobject responsiveness, the cru
cial factor in psychopathology is disturbance in the self-selfobject rela
tionship. If the selfobject responses are not optimally frustrating, that
is, if they are excessively frustrating or stimulating, the vulnerable
childhood self is threatened and will be forced to erect defenses to
protect itself. Kohut calls this threat "disintegration anxiety," the fear
of loss of self. The fear of loss of sense of who one is, according to
Kohut, is the deepest form of human anxiety; it underlies all pathol
ogy. Because disintegration anxiety is so intolerable, the self will
always choose to protect itself, no matter what the cost. Kohut called
this the "principle of the primacy of self-preservation." Protection of
the self has greater motivational power than any type of libidinal or
other gratification, and its failure is feared more than death. Con
sequently, threats to the self call forth strong defenses that protect the
weak childhood self but, in so doing, block its further development.
All psychopathology, in Kohut's view, ultimately results from
arrested self development and, therefore implies failure of the selfob
ject milieu. Differences in pathology are due to the severity and devel
opmental stage of the selfobject failure and to the secondary conflicts
that follow from such failures. With these concepts Kohut replaced the
conflict-defense model with a model of pathology based on blockages
in the development of the self. While a detailed comparison with
Winnicott's theory is not possible here, it should be noted that Kohut's
view of pathology as arrested development strikingly parallels
Winnicott's concept of pathology as blocked maturational process.
Throughout his theoretical writings, Kohut (1971,1984) maintained
that there are three categories of psychopathology: the psychoses, nar
cissistic pathology, and the structural neuroses. The first category con
sists of "prepsychological states" operating at the level of primary
narcissism. In Kohut's (1984) view, psychotics are arrested prior to the
awareness of selfobjects, that is, prior to psychological life, and so fail
to achieve a cohesive self, cannot use selfobjects or form transferences
and are not analytically accessible. Such patients, in short, lack a
self. This does not seem to be a remarkable point of view—except
that Kohut included in this category borderline patients. While he
Heinz Kohut 255
acknowledged that his clinical experience with such patients was lim
ited, he insisted that they are psychotic, once deeming them "really
schizophrenic" (Kohut, 1971), and he differentiated them from other
psychotic patients only by their ability to defend against their psy
chosis. Since patients in this category are not psychologically accessi
ble, Kohut does not advocate psychoanalytic treatment for them and
purports to offer no therapeutic contribution to this type of pathology.
In Kohut's second category of psychopathology are patients who
form a cohesive self that is enfeebled, lacking in vitality, and easily
threatened. Kohut believed that this group of patients suffers from
narcissistic disorders. In these cases the self has developed to the point
of cohesion; however, owing to failures in the parental selfobjects, the
development of this nuclear self is arrested and it becomes prone to
temporary fragmentation.
The highest level of pathology in Kohut's scheme consists of the
structural neuroses. Unlike narcissistic patients, neurotic patients
have a strong sense of self. They suffer not from an enfeebled self
but from the inability to complete the achievement of goals. In
Kohut's view, a neurosis is not a defect in the self structure, but a
product of the frustration of a self that has not been allowed to
complete its "nuclear program."
Since Kohut purported to make no contribution to the understand
ing of psychotic states and focused primarily on narcissistic pathol
ogy, the discussion of his theory of psychopathology will begin with
his views on the latter.
Narcissistic Disorders
Kohut's views on pathological narcissism emanate directly from his

developmental theory. Narcissistic personality and behavior disor
ders both result from faulty selfobject responses to the narcissistic
needs of the growing child between the phase of primary narcissism
and the oedipal phase. If there is a consistent lack of optimal frustra
tion in this phase, the child is faced with the humiliation of its
grandiose-exhibitionistic needs or a longing for idealization and
erects staunch defenses against them. The result is arrest in the devel
opment of the grandiose-exhibitionistic self and/or of the idealized
parental imago. The selfobject failure with respect to the needs of the
grandiose self stems from a lack of sufficient mirroring; selfobject fail
ure with respect to the idealized parental imago is due to the unavail
ability of suitable idealizable objects. If there is a deficit in either the
grandiose self or the idealized object, an attempt will be made to erect
256 Chapter 6
compensatory structures by strengthening the other pole of the self to

compensate for the defective pole. If this effort at compensation is not
effective because of selfobject failure in the second pole, defenses will
be used to the exclusion of compensatory structures, self development
will be arrested, and the self structure will be defective. In Kohut's
view, only the concurrence of both types of selfobject failure leads to
narcissistic fixation. However, each case of narcissistic personality or
behavior disorder tends to be characterized by a predominant defect
in one of the two poles of the self, depending on whether the greater
trauma lies in faulty mirroring or in the unavailability of the idealized
parental imago, although the other pole is sufficiently damaged that
compensatory structures cannot make good the defect.
As we have seen, in Kohut's view, the self structure is formed from
the transmuting internalization of the grandiose self and the idealized
parental imago. If either or both become arrested, the self is unable to
develop completely and remains weak, lacking in vitality, and dishar
monious. Instead of realistic ambition, there will be arrested grandios
ity; instead of tension regulation and the idealization of the superego,
there will be continuing dependency on idealized figures. In either situ
ation, transmuting internalization has either not taken place or has been
incomplete. It may be said that all narcissistic pathology, in Kohut's
view, results from the failure of transmuting internalization in the nar
cissistic phase of childhood. In accordance with Kohut's presentation of
narcissistic pathology, we will consider the failures of idealization first.
Any premature disruption in the idealized parental imago deprives
the child of the opportunity to come to see gradually the realistic qual
ities of the parent. This is most pronounced in cases of early parent
loss, but it applies to any forced relinquishment of the idealized imago
before the psyche is able to cope with it. According to Kohut, the most
potent pathogenic factor in this situation is the child's loss of the
opportunity to see the parent realistically, with the result that the
early idealization endures. The perception of the parent continues in
fantasy instead of being modified by reality, and the child is not able
to use the gradual disillusionment in the fantasied parent to build
psychic structure. Consequently, the idealized parental imago is either
repressed or split off. In either case, it endures without modification
by reality in the psyche of the child. The absence of transmuting
internalization results in a deficit in psychological structure.
Trauma with respect to the idealized parental imago may occur in a
variety of ways. The parent may be so unavailable that the child
is deprived of the opportunity to idealize the parent, or the idealiza
tion may occur but be prematurely disrupted, such as by parental
Heinz Kohut
withdrawal or loss of the parent. Whatever the source of defect in ide

alization, the result is a gap in psychic structure and a continual need
to fill this gap. Individuals who suffer from disturbances in the ideal
ized parental imago have a continual hunger for objects; the object is
not sought for its own qualities but to make good an internal defect.
This use of the object is the distinguishing feature of narcissistic dis
turbance as opposed to the idealization characteristic of the neurotic
patient. The latter will exaggerate positive qualities, but the object is
separate and perceived as a complete person, even if viewed in an
exaggeratedly positive fashion. According to Kohut, the object hunger
of the patient suffering from disturbance in the earlier idealized
parental imago has no conception of the object as a separate person,
since the object is being used to fill a gap in the self.
In accordance with his view of the dual function of the idealized
parental imago to control drives and idealize the superego, Kohut
defined three separate levels of disturbance in the transmuting inter
nalization of the idealized parental imago. The very earliest level is
the most disruptive since it leaves a gap in the psyche: only a minimal
formation of psychological structure has taken place and therefore dif
fuse narcissistic vulnerability results. Later disruptions in the ideal
ized parental imago lead to deficiencies in one of the two functions of
the idealized parental imago. If the disturbance occurs after a greater
degree of structure has formed, the gap in the fabric of the psyche
leads to a deficit in neutralizing capacity and the ability to control the
drives. If the idealized parental imago is traumatically disrupted upon
entry into the oedipal phase, the result is a failure in the idealization
of the superego. In this case, the superego may be well formed as a
result of the vicissitudes of object-libidinal strivings, but compliance
with its demands will provide little joy. Lacking gratification from the
adherence to its own internalized values, the child attempts to restore
narcissistic balance by the admiration of external figures. Such an
"idealization" is not the appreciation of admirable qualities but a
desperate effort to achieve narcissistic balance by filling a gap in the
psyche.
A good clinical illustration of Kohut's (1971) understanding of nar
cissistic pathology originating in disturbance of early idealization is
his discussion of Mr. A. The patient presented with a complaint of
homosexual preoccupation. Although he had never had a homosexual
experience, Mr. A did have occasional sexual fantasies of controlling
and dominating powerful men. During the analysis it became clear
that far more crippling than these fantasies were his tendency to feel
depressed and lethargic and his exquisite sensitivity. The transference
Chapter 6
consisted of two primary demands: that the analyst share the patient's
values and that he confirm "through a warm glow" that the patient
had lived up to the analyst's standards. When the patient did not feel
this glow from the analyst, his behavior, no matter how moral, felt
empty and trite. In the course of the analysis Kohut found the patient
to suffer from exquisite and diffuse narcissistic vulnerability. He was
easily injured by any imperfection in the analyst's ability to achieve
complete and immediate understanding of him.
In his effort to understand Mr. A's pathology Kohut first focused
on the patient's mother. He believed that his present-day behavior jus
tified the inference that she was a "deeply disturbed" woman who
easily disintegrated under pressure. Kohut (1971) concluded: "It may
thus be assumed that the patient suffered many disappointments in
the mother's phase-appropriately required omniscient empathy and
power during the first year of his life and that the shallowness and
unpredictability of his mother's responses to him must have led to his
broad insecurity and narcissistic vulnerability" (p. 61). This lack of
empathic attunement by the mother left the child with early narcis
sistic fixations because his infantile grandiosity had insufficient
opportunity to become transformed into "drive-regulating structures."
In Kohut's view, even more important to the patient's pathology
than his traumatic disappointment in the mother's lack of empathy
was the lack of sufficient opportunity to idealize the father. This in
itself may not have been pathogenic except that the boy suffered a sec
ond traumatic loss of idealization. The father had been a successful
businessman in Europe before World War II, and the family, being
Jewish, was forced to flee the Nazis, resulting in the loss of the family
business and financial status. Before that time the father and son had
been close and the boy had greatly admired his father. According to
family lore, the father had often taken his young son to the workplace
and had discussed the business with him. When the family emigrated
to the United States, the father tried many times to reestablish his
business, often with great hope for success, only to be disappointed on
every occasion. Thus, Mr. A as a boy had suffered repeated disap
pointments "in the power and efficacy of his father just when he had
(re-)established him as a figure of protective strength and efficiency"
(Kohut, 1971, p. 58). Thus, the father, onto whom all the patient's
needs for attachment to an idealized figure were placed, disappointed
his son traumatically.
Kohut concluded from this pattern of traumatic disappointment
that no transmuting internalization of the idealized parental imago
had taken place. In Kohut's view, this failure resulted in a gap in the
Heinz Kohut 259
psyche where the idealization of the superego would normally be.

Here Kohut applied his distinction between narcissistic and object-
libidinal development. According to Kohut, since the object-libidinal
attachment to the father was undisturbed, the superego was well
formed. However, since the idealization of the father did not result in
the transmuting internalization of the idealized parental imago, no
narcissistic investment in the superego—the heir of the relationship
with the father—took place. The patient had an intact superego, but
since there was no narcissistic investment in it, compliance with it
gave him no pleasure.
Lacking idealization of his superego, Mr. A needed the analyst
both to share his values and to "confirm with a warm glow" that he
had lived up to the analyst's standards in order to feel any meaning
in the achievement of his life goals. In addition, Mr. A continued to
seek replacements for the lost father of his early years by soliciting
the counsel and approval of men whom he considered superiors or
elders, despite his considerable level of success. When he felt their
approbation was not forthcoming, he felt enervated and lost enthusi
asm; it was only when he felt they expressed interest in and
approval of him that he felt complete and whole. In Kohut's view, it
is typical of the relationships formed by patients who have suffered
traumatic disappointments in the idealized parental imago that
older persons are sought out to fill the gap in the psyche left by the
traumatic disappointment in the idealized parental figure.
Kohut attributed Mr. A's homosexual fantasies to the sexualization
of the narcissistic disturbance. The fantasies involved his dominating
and enslaving physically strong men, were not always orgastic, and
sometimes did not even involve masturbation. On occasion, Mr. A
achieved orgasm with the fantasy of masturbating a strong, powerful
man and draining him of his power. Kohut interpreted these fantasies
as the sexualization of Mr. A's need for the approval of older men.
According to Kohut's formulation, Mr. A's sexual attraction to power
ful men was correlated with his demands for the analyst's approval.
Further, the fantasies of draining powerful men were manifestations
of Mr. A's need for psychological structure. He sexualized his need for
a strong, firm psyche, which was in contradistinction to his own sense
of enfeeblement. This sexualization took place, in Kohut's view,
because the early narcissistic disturbance damaged the neutralizing
capacity of the psyche; as a result, the drives were not well regulated.
This lack of neutralizing capacity and the gaps in the psyche created by
premature and traumatic disappointment in the hypercathected ideal
ized paternal imago resulted in a narcissistic disturbance that was
260 Chapter 6
characterized by homosexual fantasies; sensitivity to injury, resulting

in frequent depression; and the constant emotional dependence on
approval from older men.
Kohut understood Mr. A's homosexual fantasies, frequent bouts of
lethargy and depression, and his chronic lack of self-esteem as ema
nating from the gap in his psyche left by the too-early, too-abrupt
loss of the idealized father. He also included in his formulation of this
case a consideration of the mother's lack of empathy, which he pre
sumed led to the special importance the idealization of the father
assumed. Further, the lack of maternal empathy meant that the
mother could not be turned to in order to compensate for the distur
bance in the idealized parental imago. The case of Mr. A is a good
illustration of Kohut's views on the pathological results of trauma to
the idealizing parental imago: Kohut accounted for all Mr. A's symp
toms by the inability to idealize the superego and from the absence of
drive-regulating structures.
Having discussed Kohut's views on disturbances in the child's need
for idealization, we now turn to the pathological effects of arrest in the
other narcissistic structure, the grandiose self. If there is a defect in the
parental capacity to mirror the child's grandiosity, the child's narcissis-
tic-exhibitionistic needs are frustrated and are then either split off or
repressed, leading to the arrest of the grandiose self rather than to its
modification by reality (Kohut, 1971). Whether the grandiose self is
split off or repressed, the result is a split in the psyche; the child's
grandiosity remains out of touch with the reality-based self, rather
than becoming gradually integrated into it.
If the grandiose self is repressed, the result is general psychological
impoverishment, including low self-esteem, vague depression, and
lack of initiative. If the grandiose self is split off, the psyche then has a
structure that relates to reality and a grandiose self that functions but
is walled off, without contact with the rest of the psyche. The split-off
grandiosity is manifested in a boastful, superior attitude, sometimes
with cold arrogance. Kohut (1971) calls this a "vertical split," which he
differentiates from the "horizontal split" caused by repression. Since
the narcissistic cathexes are invested in this split-off grandiosity, the
more realistically oriented part of the self is robbed of self-esteem.
Consequently, the conscious feelings tend to be lethargy, depression,
and emptiness, similar to the symptomatic manifestation of repressed
grandiosity. The difference is that in vertically split off grandiosity the
psychological impoverishment coexists with the arrogant, haughty,
superior attitude. In addition, the empty, lethargic self represses the
vulnerable nuclear self to protect it from feared injury. As a result, the
Heinz Kohut 261
usual structure of pathological narcissism includes both a vertical split

between grandiosity and the reality-oriented empty self and a hori
zontal split in which the depleted self represses the vulnerable nuclear
self with its unfulfilled narcissistic longings.
Whether the grandiosity is split off or repressed, the psyche is
robbed of its primary source of self-esteem. Without this supply of
narcissistically invested libido, self-esteem is minimal, and the child
will develop a hypersensitivity to criticism and lack of approval.
When deficits or limitations of any type are apparent, the personality
reacts with shame or rage. There is no solid source of self-esteem with
which to combat injuries, whether real or apparent. It is not simply
that such patients were not mirrored and so do not value themselves.
In Kohut's view, the lack of mirroring results in the need to protect the
infantile grandiosity with a massive defense so that the personality is
deprived of its source of self-esteem. The archaic grandiosity does not
become transformed into realistic self-esteem, and the resulting per
sonality is characterized by both archaic grandiosity, with the need to
be mirrored constantly, and a poverty of self-esteem. The arrested
grandiosity is walled off from the rest of the psyche, robbing it of its
most important source of self-esteem, and thereby leaving it vulnera
ble to narcissistic injury. This formulation is Kohut's explanation of
the apparent contradiction that patients suffering from pathological
narcissism are grandiose, feeling superior to others, and yet are so
vulnerable that they have an exquisite sensitivity to slights.
One can see that Kohut's view of pathological narcissism is clearly
in conflict with Kernberg's conception of the disorder (discussed in
chapter 5). The principal difference is that in Kernberg's view, the
grandiose self and primitive idealization are both defenses, pathologi
cal constructions to protect against oral rage, envy, dependence,
emptiness, and paranoia. For Kohut, pathological narcissism, whether
due to a fixation of grandiosity or idealization, is a developmental
arrest; the longings of the patient are "phase appropriate" for the nar
cissistic stage of development. As we shall see in the next section, this
difference leads Kohut to a sharply different treatment approach.
The vertical and horizontal splits means that the grandiose-exhibi
tionistic needs of the self are unable to find expression. Consequently,
there is a buildup of tension due to the dammed-up exhibitionism. In
Kohut's view, typical expressions of this unconscious exhibitionism are
hypochondriasis, a propensity to shame, and self-consciousness. These
symptoms are all the products of excessively repressed or split-off
exhibitionism that overtaxes the capacity of the psyche, and all three
are common to narcissistic personality disorders. Unable to contain the
262 Chapter 6
push for the expression of the exhibitionistic-grandiose needs, the psy

che expresses its tension somatically, resulting in hypochondriacal
symptoms. The tendency to feel shame comes from the lack of self
esteem in the realistic portion of the psyche: any limitation is experi
enced as a revelation of the inadequacy of the self. Similarly, the
constant fear of being exposed gives rise to continual self-conscious
ness. These symptoms tend to characterize the narcissistic personality
disorder whose deficit lies primarily in the arrest of the grandiose self.
Nonetheless, one can find similar symptoms in many neurotic
patients. Kohut warns that one cannot make a diagnosis of narcissistic
pathology on the basis of symptoms. It is not the hypochondriasis,
self-consciousness, or shame in and of themselves that warrant the
diagnosis of narcissistic pathology. The decisive pathognomonic fea
ture is the source of the symptoms, the split-off grandiosity. The key
to whether the presenting symptoms are due to walled-off grandiosity
is quality of the object relationships. If the patient treats others only to
suit his or her own purpose, one can safely conclude that the underly
ing pathology is narcissistic, which means that any of the symptoms
being discussed is a product of split-off grandiosity. If others are
treated as full persons in their own right, the symptoms are most
likely due to a structural neurosis.
In addition to shame, the other major reaction of the narcissistically
vulnerable patient to disappointment and injury is narcissistic rage. In
Kohut's (1972) view, aggression is most destructive not when it is
impulsive but when it results from narcissistic injury. The threat to the
self and the sense of shame lead to a need to inflict injury on the perpe
trator, and the revenge motive knows no rest until the injury is repaired
by satisfaction of the need to avenge the wrong. In Kohut's view, this
compulsion to repair injury by inflicting it differentiates narcissistic
rage from other forms of aggression. Because the motive is repair of
narcissistic injury, the rage and revenge show no regard for reasonable
human limitation in their implacable need to eradicate the injury. The
irrational pursuit of revenge and repair reflects the narcissistic motiva
tion of the aggression. Other types of aggression, such as competitive
hostility, do not include the compulsive need to avenge injury. Patients
suffering from arrested archaic grandiosity are easily injured and are
prone to bouts of rage whenever their expectations for mirroring are
not met. For this reason, narcissistic patients are given to chronic rage
attacks. Indeed, the same holds for patients suffering from arrested
development of the idealized parental imago. When the need for ideal
ization is frustrated, the self is threatened and the patient will become
enraged and seek revenge on the source of the threat.
263
In accordance with his theory of development, Kohut (1971) delin

eates three levels at which the grandiose self may become arrested. At
the lowest level so little psychological structure has taken place that
others are experienced as part of the mind or body of the grandiose
self. If a greater degree of self development has occurred, others are
perceived as like the self but not a part of the self. If development is
arrested closer to the oedipal phase, others are experienced as sepa
rate but are significant only insofar as they fulfill the needs of the
grandiose self for admiration and mirroring. At all three levels, others
are experienced as selfobjects; that is, they exist, from the point of
view of the patient, only insofar as they meet a narcissistic need at one
of these levels. In short, the other is not recognized as a separate per
son. When narcissistic needs are not fulfilled at any of these levels, the
self is threatened and the patient feels a profound sense of shame or
rage or both.
Kohut's (1968) view of arrested grandiosity is illustrated by Miss F,
who sought analysis because she was unable to be intimate and felt
"different from other people and isolated" (p. 503). In the analysis it
became clear without apparent cause that she was given to abrupt
changes between anxious excitement and elation, when she experi
enced herself as "precious" and consequently felt superior to others,
and states of emotional depletion. In the analysis Miss F demanded
that the analyst repeat to her what she said but not go beyond it. When
Kohut complied, she became calm and content, but when he either
remained silent or attempted to add to her comments, she became
enraged, feeling misunderstood and undermined. During her child
hood her mother had been depressed during the narcissistic phase, and
Kohut concluded that her grandiosity had not been mirrored, resulting
in the splitting off of the grandiose self from the rest of the personality.
Consequently, Miss F attempted to use others for the narcissistic echo
she missed in childhood. Kohut began to see that he was not a sepa
rate, autonomous person to her but a function. In her view, Kohut's
role was to fill the gap in her psyche by mirroring her comments.
When he did so, she felt the elation of having her grandiosity, her "pre
ciousness," confirmed; when he did not, she felt the depletion of unful
filled grandiose-exhibitionistic needs. This constellation, in Kohut's
view, is characteristic of narcissistic personality disorders. The arrest
of the grandiose self results in depletion of the psyche and a need for
others to approve of and confirm the grandiosity in order to fill the
deficit in the self. Miss F's isolation and inability to be intimate were a
direct result of her perception of others not as whole persons but as
impersonal functions for the completion of her self.
264 Chapter 6
These same dynamics of arrest in the grandiose self and idealizing

parental imago apply to narcissistic behavioral disorders. The differ
ence is that in the narcissistic personality disorders the weakness in
the self results in symptoms such as hypochondriasis, depression, and
lethargy; in the narcissistic behavioral disorders the patient attempts
to fill the narcissistic deficit with action, such as delinquency, perver
sions, or addictions (Kohut and Wolf, 1978). Both are disorders of the
self in that the self is enfeebled and lacking in cohesiveness and vital
ity, but in one case the weakness in the self is directly expressed in
symptoms whereas in the other the patient takes action to fill the gap
so that the weakness of the self is not experienced. Although Kohut
included delinquency in his list of narcissistic behavioral disorders,
his discussion of the dynamics of this type of narcissistic pathology
was confined to addictions and perversions. We will approach the
addictions first, focusing on drugs and eating disorders, and then use
the fetish as an example of his understanding of perversions.
Kohut (1971) viewed drug addiction as a self disorder in which the
drug is used to fill the missing gap in the psyche. According to Kohut,
the mother of the addict failed to perform her tension-regulating and
other functions, resulting in a traumatic disappointment in the idealiz
ing selfobject. Consequently, the child is robbed of the opportunity to
internalize the object gradually to form psychological structure. The
addict uses the drug not so much as a substitute for what was not
given, such as love or food, but to fill the gap left by the missing ideal
ized object. Whereas patients with narcissistic personality disorder
attempt to form idealizing attachments to people to fill the gap in their
psyche, drug addicts, representing a form of narcissistic behavioral
disorder, use the drug in a desperate effort to achieve the same goal.
Kohut (1977) viewed eating disorders in a similar fashion. He did
not believe that pathological overeating is a result of strong oral dri
ves or fixation at the oral level. In his view, the child needs a "food-
giving selfobject." If the environment does not meet this need, the
child does not feel responded to as a whole self and regresses to
intensified, fragmented pleasure-seeking oral stimulation. The result
is an addiction to food. Pathological eating is a good illustration of
Kohut's belief that the drives intensify and become fragmented in
response to disruptions in the self. The selfobject failures traumatize
the grandiose self and lead the child away from its center of initiative
and toward pleasure seeking to feel whole and real. Overeating is an
effort to experience the feeling of wholeness without relying on a
failed and untrustworthy human environment.
Kohut applied a similar analysis to fetishes. In his view, the fetish is
Heinz Kohut 265
a substitute selfobject that patients use to attempt to receive the sooth

ing they did not get from the mother. Faulty maternal empathy leaves
the grandiose self unresponded to, and the child regresses to "archaic
pleasure gains" in response. The child finds a substitute that will pro
vide the soothing merger feeling he or she could not obtain from the
human environment. The fetish is not so much a substitute for food or
love as a desperate effort to fill a defect in the self; the fetishistic object
substitutes for the missing admiration and approval needed for the
nourishment of the grandiose self.
Kohut's (1977) understanding of the fetish is illustrated by the
case of Mr. U, a brilliant but relatively unsuccessful college profes
sor who received little enjoyment from his work and entered analy
sis to eliminate his need for certain fetishistic objects, such as nylon
stockings and underwear. His previous analyses had focused on
castration anxiety and denial. Kohut, however, believed the fetish
was related to Mr. U's "oddly empathic, unpredictable, emotionally
shallow mother" (p. 55) who would become overly involved with
him, caressing him and meeting his wishes, and then suddenly
withdraw and become unresponsive. Kohut believed that Mr. U
withdrew from the unpredictability of his mother to the soothing
touch of objects, which eventually developed into the fetish. The
father was self-absorbed, and all the patient's efforts to be close
were rebuffed. According to Kohut, the boy was deprived of the
opportunity to idealize an object and gradually relinquish it. The
lack of internalization of an idealized object left him without inter
nalized ideals and solid self-esteem. He felt either inadequate to
achieve lofty ideals or superior to worthless goals, resulting in
abrupt and severe shifts in self-esteem. In a desperate effort to boost
his feeling of self-worth, the patient intensified his attachment to
the mother-fetish. This case is a good illustration of Kohut's views
of the fetish and of narcissistic behavioral disorders in general.
Whereas a narcissistic personality disorder would demand attune-
ment and admiration and become ashamed or enraged at not receiv
ing it, Mr. U attempted to achieve his sought-for merger through his
attachment to the fetish.
As can be seen from this discussion, Kohut was able to explain a
wide range of pathology by failures in the self-selfobject relationship
and the resultant arrest in development of the self. He explained
symptoms and disorders in the character disorder range on the basis
of failure in narcissistic development. Symptom syndromes, such as
depression and hypochondria, and behavioral syndromes, such as
perversions, addictions, and delinquencies, were all explained as
266 Chapter 6
manifestations of a depleted self. The question of Kohut's understanding

of neurosis now arises, and it is to this subject that we now turn.
The Structural Neuroses
Kohut's position on the structural neuroses shifted in accordance

with the causal role he attributed to the Oedipus complex. When
Kohut limited his theoretical contribution to narcissistic disorders,
he believed that the development of the self was completed before
the oedipal phase, resulting in the separation of self from object. In
this view, the child has moved from the selfobject phase, in which
objects serve functions for the self, to the oedipal phase, in which
others are seen as separate, autonomous objects who become the tar
get of aggressive and sexual feelings. Thus, in this formulation, nar
cissistic disorders are products of faulty self development, and
conflict, occurring in the oedipal phase, is responsible for structural
neurosis. That is, in his early theory Kohut attributed neurosis to
unresolved oedipal conflict. However, Kohut changed his view of
the importance of the oedipal phase as he developed a broader view
of self psychology. The logical consequence of this shift was a modi
fication of his views of the dynamics of structural neurosis. Kohut
(1984) came to see oedipal conflicts as a product of a weak, defective
self. If the parents are threatened by the child's assertiveness or
affection and respond with competitive hostility or overstimulation,
the child experiences a traumatic disappointment in the oedipal self
objects, and the development of the assertive, affectionate self is
arrested. The joyful affectionate and assertive feelings become inten
sified and are split off from the core of the self and transformed into
gross sexuality or hostility. It is in this way that the oedipal phase
becomes pathological, and it is to this pathological constellation that
Kohut gives the name Oedipus complex.
In Kohut's view, the primary fear of the girl in the oedipal phase
is a seductive father or a hostile mother; the boy fears a seductive
mother and hostile father. If these fears are realized, selfobjects
become threatening rather than empathic and the self begins to disin
tegrate; to protect itself from spreading anxiety, the child isolates the
drives from the rest of the self. At this point, affection becomes sexual
ity and assertiveness hostility. These sexual and hostile responses,
according to Kohut, always serve to protect threats to the self. The
nuclear self, in his view, includes affectionate and assertive feelings,
but if these feelings are threatened by traumatic selfobjects, it will pro
tect itself by isolating these positive feelings and transforming them
267
into sexuality and hostility. Since the isolation of the drives protects
from a threat to the self, the Oedipus complex is ultimately an effort to
defend against "disintegration anxiety."
Kohut (1984) does acknowledge that since the same-sex parent is
now hated and feared, the child will feel a threat from its rival, which
boys experience as castration anxiety. Kohut does not discount the
phenomenon of castration anxiety, but in his view castration anxiety is
not a cause of symptoms, but symptomatic of a threatened self. Just as
the hostility to the same-sex parent is self-protective, so too is its
sequel, castration anxiety: the focus on loss of the penis defends
against the far more threatening loss of the integrity of the self.
According to Kohut (1971), classical analytic theory does not analyze
castration anxiety to its root; the biological bias of classical theory
leads it to the assumuption that castration anxiety is psychological
bedrock, but this focal anxiety always masks a far more frightening
anxiety, the threat to the integrity of the self. In Kohut's view, psycho
logical bedrock is the cohesiveness of the self and the deepest anxiety
is any threat to the integrity of the self.
In Kohut's reformulation of the oedipal phase, the Oedipus com
plex is a pathological product of faulty selfobject responses to the
oedipal-age child. Since oedipal conflicts result from a defect in the
self, we are left with the question of whether one can do away com
pletely with the concept of structural neurosis and instead adopt the
view that all neuroses are really narcissistic personality disorders.
Kohut (1984) rejected this conclusion and, in fact, in his last work
defended the concept of structural neurosis, including the term trans
ference neurosis, but he distinguished between the structural weakness
of the self and the developmental failures that lead to those weak
nesses. If such weaknesses occur in the preoedipal phase, they lead
either to perversions and addictions, as in narcissistic behavioral dis
orders, or to the symptoms of an enfeebled self, as in narcissistic per
sonality disorders. If they occur in the oedipal phase, they lead to
intensification and isolation of the drives; the drive fragments must be
defended against, and neurosis is the outcome.
According to Kohut's final view, both pathological narcissism and
neurosis are the product of faulty selfobjects, resulting in defects in
the self. They differ in the developmental phase in which the self
defect occurs and therefore in the resulting symptoms. In pathological
narcissism the nuclear self is enfeebled, lacking in vitality, and dishar
monious; in structural neurosis the nuclear self is unable to reach its
goals not only because its development has been arrested by the lack
of mirroring response by the selfobjects of the oedipal phase but also
268 Chapter 6
because its energies become absorbed by these oedipal conflicts, so

that it is unable to realize its ambitions and ideals.
Kohut (1984) illustrated the self-psychological approach to the neu
roses in his discussion of agoraphobia. According to the classical view,
the agoraphobic woman cannot leave the house because of the incestu
ous wish that is displaced onto the males she might see on the street.
According to Kohut's reformulation, the faulty selfobject milieu in the
oedipal phase is the primary cause of the disorder. This is so for two
reasons. First, the failure of the paternal selfobject to respond apprecia
tively to the little girl's affection and assertiveness as she enters the
oedipal phase results in the breakdown of the self and the conversion
of these positive attitudes into lust and hostility; that is, the advent of
the Oedipus complex results in the disintegration of normal affection
for the father and its conversion to sexual fantasies. Second, the faulty
maternal selfobject results in failure to internalize self-soothing func
tions, a failure that leads to the tendency for anxiety to become con
verted into disintegration anxiety. The patient has developed no
self-soothing functions and therefore cannot leave the house without a
maternal figure to soothe her; leaving the house alone threatens an out
break of disintegration anxiety. Kohut (1984) concluded that "the
addictionlike need for an accompanying woman is not to be viewed as
a defensive maneuver but as a manifestation of the primary disorder:
the structural defect of which both the unconscious oedipal fantasy
and the conscious need for a female companion are symptoms" (p. 30).
In Kohut's view, agoraphobia is illustrative of the pathogenicity
of faulty self-selfobject relationships. Moreover, neuroses, each with
its own specific dynamics, involve the failure of the oedipal selfob
jects, and it is this failure that results in the breakdown of the self
with the threat of disintegration anxiety and the onset of the patho
genic Oedipus complex. The neurotic symptom, then, is an effort to
bind disintegration anxiety; thus, in each neurosis the intensity of
the drives is symptom, not cause. Thus, we see that Kohut does give
the oedipal phase an intermediary role in the onset of neurosis. Once
the selfobject milieu fails in the oedipal phase, the weakened self
engages the Oedipus complex, with the ensuing conflicts around lust
for the opposite-sex parent and hatred for the same-sex parent,
resulting in neurotic symptoms, but, for Kohut, the primary cause of
these symptoms remains the defect in the oedipal self that results
from faulty selfobjects.
With this view of neurosis, Kohut (1984) adopted the view that
all psychopathology is based on disturbance in the structure of the
self. As we have seen, his initial position was that severe pathology
Heinz Kohut 269
involves defects in the self whereas structural neuroses result from

conflicts that arise after the complete formation of the self; then, as he
began to expand self psychology, he ultimately concluded that incom
plete self development is the primary cause of neurotic pathology, too.
We have seen that Kohut maintained a distinction between pathologi
cal narcissism and neurosis despite his belief that all psychopathology
consists of defects in the structure of the self. Does this view of neuro
sis as self-pathology mean that the drive-ego model of the structural
neuroses can be dispensed with in favor of the self-psychological
view? Kohut (1984) himself raised this question and stated that he
could not give a definitive answer although in the future the answer
might well be in the affirmative. However, this tentativeness is in
apparent conflict with his analysis of neurosis as self pathology, and
his views on oedipal dynamics and the role of the drives. If neurosis
involves a primary defect in the self, then it is a form of self pathology
and it follows logically that its treatment, to be discussed shortly,
would follow the guidelines of self psychology. In Kohut's scheme no
primary causal role is attributed to the drives or to the Oedipus com
plex; thus, the drive-ego model has, in fact, been replaced by the self
psychology model (despite Kohut's unwillingness to acknowledge
this). Indeed, we saw an example of this in Kohut's discussion of ago
raphobia, in which he delineated the self-psychological view of this
neurosis: he indicated clearly his view that self psychology offers an
alternative to the drive-ego way of understanding the disorder.
At this point it becomes clear that Kohut has broadened self psy
chology into a model for understanding all pathology, thereby replac
ing the traditional emphasis on the vicissitudes of the drives and
defenses against them with an emphasis on the structure of the self.
According to Kohut, pathology is arrested development, not conflict
within the self, and defenses are efforts at self-preservation that play
no causal role in pathology. The source of all pathology is not conflict
ing forces within the self but developmental arrest rooted in failures
in the selfobject environment. The relationship between self and self
object is the crucial variable for understanding such failures and how
they eventuate in psychopathology. This new psychoanalytic para
digm of psychopathology has clear and far-reaching implications for
treatment, and it is to these implications that we now turn.
TREATMENT
Having reformulated psychopathology as developmental arrest due

to selfobject failure, Kohut viewed the psychoanalytic process as the
270 Chapter 6
means by which the arrested self is able to complete its development

with a new selfobject experience. This process consists of the thera
peutic mobilization of the arrested self, the use of the analyst as the
selfobject that was missing in development, and the transmuting
internalization of this selfobject into psychological structure. The ana
lytic process is based not on the interpretation of defense to uncover
drive wishes but on the overcoming of developmental arrest. The
analytic situation, if not disrupted with premature interpretations or
analytic "moralism," allows the spontaneous therapeutic mobilization
of the arrested self (Kohut, 1971). The analyst empathically immerses
himself or herself in the patient's experiential world through accurate
and well-timed interpretations. This analytic empathy allows the
analyst to be used as a selfobject. Thus, interpretations are key to the
analytic process not because of the knowledge they provide but because
they convey empathic attunement with the patient's experiential
world. Failures in the analyst's empathy are disruptive to the self
selfobject relationship but also provide opportunities for the transmuting
internalization of the selfobject into psychological structure.
Kohut divided the interpretive process into two phases. In the under
standing phase, the analyst grasps the meaning of the patient's behavior
and communication. For example, if a patient feels disrupted by the
analyst's approaching vacation, the analyst's connection of the patient's
affective state with the interruption in treatment demonstrates his or
her understanding to the patient. In the explanatory phase, the analyst
looks for the genetic-dynamic determinants of what has been under
stood. To continue our example, the analyst's explanatory communica
tion to the patient might correlate the patient's reaction to the vacation
with a childhood experience and its current influence. In Kohut's view,
understanding must always precede explanation. Kohut felt that classi
cal analysts tend to rush too quickly to explain without first under
standing. The understanding phase conveys the analyst's empathy,
making possible the reexperience of the developmental arrest, which
sets the stage for the transmuting internalization that did not take place
in childhood. To explain without understanding is to demonstrate a
lack of empathy with the patient's experience and therefore to re-create
the empathic rupture of childhood rather than help to repair it. Under
standing is a crucial component of the empathy needed for the analytic
completion of the developmental process; for some patients it may con
sume a great deal of the analysis. As a corollary to this two-phase view
of the analytic process, Kohut (1984) advocated interpretation of trans
ference first, with genetic reconstruction to be employed only after the
transference is understood.
Heinz Kohut 271
Since Kohut viewed borderline and psychotic conditions as syn

dromes in which no self has developed and therefore as inaccessible
to analytic intervention, he professed to offer no contribution to the
treatment of those forms of psychopathology. He initially offered his
specific treatment views within self psychology in the narrow sense
by focusing on understanding narcissistic disorders while he still
believed in the application of the drive-ego model to neuroses.
Eventually, as self psychology expanded beyond the borders of narcis
sistic pathology, Kohut broadened his goals to include a reconceptual
ization of the psychoanalytic process, and at that point his theoretical
shift away from traditional analytic theory began to have relevance to
the neuroses. In accordance with the development of Kohut's thought,
we will first discuss Kohut's conception of treatment for narcissistic
disorders and then consider his revised model of psychoanalysis, as he
ultimately elaborated it, and at that point, his views on the treatment
of neurotic pathology will be discussed.
Narcissistic Disorders
Since the causes of narcissistic disorders lie in arrests in the devel

opment of the bipolar self, the transferences of these patients will be
formed along the axis of either the grandiose self or the idealized
parental imago, depending on which pole of the self was more dam
aged. While Kohut fully acknowledged that some patients move
between these two types of transference, he believed that the pre
dominant transference paradigm will coalesce at one pole of the self
and that at any one time one of these two types of transference will
be in operation. Patients who suffer primarily from an arrest in
grandiosity will eventually form a mirror transference based on the
archaic need to have their grandiosity mirrored. Patients whose pri
mary defect is the arrested need to idealize a parent will form an
idealizing transference.
In Kohut's view, the analytic setting provides the conditions for
narcissistic patients to form one of the two types of selfobject transfer
ence, which will form spontaneously if allowed to do so. In Kohut's
formulation, as in the classical model, transferences are believed to be
mobilized by the regressive features of the analytic setting. The ana
lyst's task in the first phase of treatment is not to interpret the trans
ference but to allow it to unfold. Any educational measure designed
to "tame" a selfobject transference runs the risk of forcing it to be sub
merged under a wall of defenses to protect the patient's narcissistic
vulnerability. Premature interpretation has the same potential danger,
272 Chapter 6
as the infantile self is vulnerable and interpretation is likely to injure

it. The crucial technical principle in the understanding phase of the
analysis, which may last for an extended period, is that the analyst's
task is to accept the selfobject transference and understand it as a
"phase-appropriate" development. The unmet childhood needs are
now seeking gratification from the analyst, and the analyst's attitude
should be to welcome this as the therapeutic activation of the child
hood self that was not allowed to complete its development. This view
is a direct application of Kohut's belief that narcissistic transferences
represent arrested development of the self.
The acceptance of the phase appropriateness of the selfobject trans
ference should not be confused with its gratification. Kohut was quite
clear and consistent that efforts to gratify the patient's narcissistic
longings are in error because they do not allow for "optimal frustra
tion," which, as we shall see, is the crux of the therapeutic action of
the analytic treatment of narcissistic pathology. Attempts at gratifica
tion may impede the therapeutic activation of the transference, just as
rejecting responses may. Consequently, the analyst, in Kohut's view,
should foster the development of the selfobject transference but not
gratify it. In this way, the infantile self in its fullest form is most likely
to become therapeutically activated.
The only interpretations Kohut considers appropriate in this phase
of the treatment are geared toward removing defenses against the nar
cissistic transference. Some patients attempt to protect themselves
against narcissistic vulnerability by disavowing or denying their
needs for mirroring or idealization. In that case, the analyst interprets
the defenses in order to foster the development of the narcissistic
transference. However, once the transference has been mobilized, the
analyst allows its continuance without interpretation. In this phase,
interpretation must be in the service of fostering the development of
the narcissistic transference, not in resolving it. The implication is that
analysts must understand and accept that they are a selfobject for the
patient. They may well experience a narcissistic injury at being "used"
in the service of the patient's self development, as opposed to being
allowed to fulfill their professional ideal as they conceive it. The ana
lyst's ability to contain this countertransference reaction is a crucial
component of the process and is a major aspect of some variants of
self psychology, as will be discussed later.
These treatment principles apply equally to both types of nar
cissistic transference. All further treatm ent recommendations
are specific to the type of narcissistic deficit mobilized in the trans
ference. In accordance with Kohut's presentation, we will discuss
Heinz Kohut 273
the idealizing transference first and then the mirror transference.

In the idealizing transference the patient uses the analyst to fill the
gap in the psyche by accepting the analyst either as a drive regulator
or as an external figure to complete the idealization of the superego. In
either case, the attachment to the analyst becomes essential to the
functioning of the patient's self. By means of this attachment the
patient is able to function, but the fact that an external figure is needed
indicates that the self is defective. The distinguishing feature of the
narcissistic, as opposed to the neurotic, idealizing transference is the
utilization of the analyst as a selfobject to make good the self's func
tional deficits. The neurotic patient idealizes by exaggerating positive
qualities of the analyst who is seen and treated as a separate person,
that is, an object, rather than a selfobject. The narcissistic patient who
forms an idealizing transference uses the analyst to achieve narcissis
tic equilibrium and therefore treats the analyst as a part of the self,
rather than as a separate person.
There are three types of idealizing transference corresponding to
the levels of rupture in the idealized parental imago. If the trauma
was very early, the patient's diffuse narcissistic vulnerability results in
a global use of the analyst to achieve narcissistic balance. The idealiz
ing transference in this case is the sole means of the patient's self
esteem regulation; without it the psyche is vulnerable to narcissistic
injury of almost any type with minimal provocation. If the trauma
was somewhat later, the idealization of the analyst is used to perform
the neutralizing function missing from the patient's psychological
structure. In this case drive wishes and impulses, normally controlled
by the idealization of the superego, are controlled by the idealization
of the analyst. If the phase of trauma to the idealized parental imago
was later still, patients do not idealize their superego but instead
search for external figures who will approve of and admire their val
ues. The analyst becomes such a figure for the patient, and this use of
the analyst defines the highest level of idealizing transference.
Kohut illustrated his view of the function of the idealizing transfer
ence and his principles of its treatment in his discussion of Mr. A who
suffered from lethargy, depression, a diffuse sensitivity to narcissistic
injury, and a preoccupation with homosexual fantasies of draining
men of their power. When he received approval from men he func
tioned well, but when confirmatory responses were not forthcoming
he felt at first depleted and depressed and then became cold and
haughty. The transference consisted primarily of two demands: that
the analyst share the patient's values and that he confirm "with a
warm glow" that the patient had lived up to them. When Mr. A felt
274 Chapter 6
that his wishes were gratified, he felt whole and full of energy. When
he did not feel that the analyst shared his values or that the analyst
felt he was living up to them, Mr. A felt disappointed and depleted;
lost his zest; and eventually withdrew into a cold, haughty attitude
toward the analyst. The transference was thus characterized by "reac
tive swings toward a hypercathexis of the grandiose self" (Kohut,
1971, p. 68). At these times the patient became emotionally cold,
adopted grandiose schemes, and was hypochondriacally preoccupied.
Some of the interpretive work involved the drawing of parallels
between Mr. A's demand for approval of his values by the analyst and
his fantasies of pursuing powerful men; between his reactive
grandiosity and the "princely young men" by whom he was sexually
stimulated; and between the orgiastic experience of gaining strength
in fantasy by draining it from strong men and his need to acquire psy
chological functions to fill the gap in his psyche. The homosexual fan
tasies were ultimately interpreted as the sexualization of his need to
gain strength from an idealized figure. That is, Mr. A achieved narcis
sistic balance by sexualizing his need for idealization. Kohut (1971)
contends that "the direct interpretation of the content of the sexual
fantasies is not an optimal approach in the analysis of such cases, and
that it should at first be demonstrated to such patients that the sexual
ization of their defects and needs serves a specific psychoeconomic
function, i.e., it is a means for the discharge of intense narcissistic ten
sions" (p. 72). Kohut questioned the value of entering into the retro
spective investigation of sexual content even after this function is
shown to patients as they are already beyond the sexualization of their
conflicts at that point.
The case of Mr. A demonstrates the value of addressing the defect
in idealization in a case of presumed sexual dysfunction. Once the
narcissistic tensions were dealt with via the idealization need, Mr. A's
homosexual preoccupations abated. What appeared to be a sexual
problem was, in fact, according to Kohut, a case of severe narcissistic
imbalance due to the defect in the idealized paternal imago; thus, the
therapeutic action of the case involved interpretation of the narcissis
tic defect rather than a direct discussion of the "sexual problem." The
patient had attempted to fill the gap in his psyche with his fellatio fan
tasies, but they provided only temporary relief and had to be repeated
continually. Kohut (1977) pointed out that "the successful filling in of
the structural void could, however, ultimately be achieved in a non-
sexual way via working through in the analysis" (p. 127). Similarly,
when new structures were laid down by the internalization of the ana
lyst, Mr. A's self-esteem solidified and his other symptoms abated.
Heinz Kohut 275
The transference, then, consisted of the patient's use of the analyst

to fill a gap in the self, in this case the idealizing imago; the ana
lyst's task was to allow this process to occur and later to interpret
the narcissistic defect that the transference was filling. The working
through of the patient's use of the analyst in this way constituted a
resolution of the narcissistic pathology, and the symptoms were
relieved as this occurred.
This use of the analyst implies that any disappointments in the ide
alization during treatment will be threats to the cohesion of the self.
Such disappointments will inevitably occur, and Kohut differentiates
between major and minor "failures" of the analyst. The latter include
day-to-day failures such as misunderstanding, not being perfectly in
tune with the patient, changing appointment times, or beginning a
session a few minutes late. Major "failures" are vacations or extended
interruptions of any type. Both types of failure, or disappointment,
disrupt the patient's narcissistic equilibrium because the patient's self
esteem is based on his or her attachment to the idealized object. While
the rupture may seem trivial to an external observer, it is traumatic to
the patient because it interferes with narcissistic equilibrium, which is
delicately balanced on the idealization. Once the idealization is dis
turbed, the self loses its self-esteem and power, resulting in disintegra
tion anxiety. The consequence is a collapse of the self, which will
manifest itself as lethargy, a sense of worthlessness, and a feeling of
powerlessness.
Such disappointments in the idealization of the analyst are not only
inevitable but are, according to Kohut (1971), an essential part of the
therapeutic action with the narcissistic patient who forms an idealiz
ing transference. Since the collapse experienced by the self requires
interpretation to be repaired, it is at the point of disruptive disap
pointments in the idealization that interpretive work begins in
earnest. The analyst's role is to discover the cause of the patient's
sense of disruption and renewed symptoms. By empathic understand
ing of the patient's iatrogenic symptoms, the idealization is restored,
as the patient now feels understood by the analyst. However, the
reestablishment of the idealization is never totally complete: each dis
appointment in the analyst leads to a microscopic bit of internalization
of the functions provided by the idealized figure, and each such bit
adds to the analysand's investment in his or her own set of standards.
It is this bit-by-bit cycle of idealization, disappointment, interpreta
tion, and restoration with microscopic internalization that is the
essence of the structure-building process with such patients. To put
this another way, the essence of the therapeutic action of the analytic
276 Chapter 6
treatment of the idealizing transference is the transmuting internaliza

tion of the self-sustaining functions of the idealized analyst into a
structure of ideals and values. In this way, according to Kohut, the
analytic process completes a developmental process that was arrested
in childhood.
It is crucial in Kohut's view that the interpretations of the patient's
intense disappointments in the analyst are stated in a way that puts
the focus on the analyst's "failure." Since the transference replicates
the patient's functional reliance on the parental selfobjects, it follows
that the patient's disappointments in the analyst are a replica of the
failures of the selfobjects of childhood. If the interpretation focuses on
the patient to the exclusion of the "failure" of the analyst, there is real
danger of repeating the childhood trauma by "blaming" the patient
for the "failure" of the parent/analyst. According to Kohut, the ana
lyst must recognize that although the precipitant of the patient's dis
appointment may seem trivial, from the patient's point of view the
analyst has failed. The adoption of this stance is an empathic attune
ment with the patient because it accepts the latter's point of view as
valid. Once patients see that the analyst accepts their point of view,
they are able to re-idealize the analyst. Again, each time this occurs, a
bit of the former idealization has, via transmuting internalization,
become internalized into a bit of psychological structure.
This process is illustrated in the reaction of Mr. G to Kohut's (1971)
announcement of a vacation. The patient responded by withdrawal
from idealization to a "primitive form of grandiose self." However,
Mr. G was reacting not to the separation itself but to Kohut's defen
sive and unempathic tone of voice in making the announcement.
Kohut fully acknowledged that he had thought primarily of bracing
himself against the "coming storm" when telling the patient of his
absence:
It was in reaction to this attitude that the patient had experienced a trau
matic disappointment in my empathic capacity which he had previ
ously idealized as limitless, and no progress was made until I could
offer my understanding and thus again enable the patient to recathect
the idealized selfobject [pp. 93-94).
This vignette demonstrates Kohut's principle that the interpretation

of such disillusionments must include full acknowledgment by ana
lysts of the role they play in evoking the response. More critically, it
demonstrates the ease with which patients suffering from narcissistic
deficits are injured by seemingly minor slights, such as the defensive
Heinz Kohut 277
tone of voice in which Kohut announced his vacation to Mr. G. In

Kohut's view, the therapeutic action of the idealizing transference
consists of a myriad of such disappointments coupled with empathic,
interpretive responses of the analyst that repair the idealization, as
occurred in the case of Mr. G when Kohut realized that his voice tone
was the source of the patient's injury. It was Kohut's empathic under
standing of the patient's reaction to his tone of voice that turned the
patient's regressive reaction into a therapeutic process. Mr. G could
then not only restore his idealization but also accept that Kohut's
empathy was not limitless, thereby allowing a bit of transmuting
internalization to take place.
The working-through process of the idealizing transference occurs
primarily by the interpretation of its disruption. Resolution is possible
because, according to Kohut (1984), interpretation itself provides the
optimal frustration required for the structure building that was
missed in childhood. The empathy of the interpretation provides the
patient with a needed developmental experience, but the fact that it
is interpretation rather than gratification is also frustrating. In the
vignette of Mr. G's reaction to Kohut's vacation, Kohut's empathy
consisted of his understanding of the patient's reaction to his voice
tone, but Kohut did not attempt to gratify the patient by apologizing
for his empathic lapse, or by canceling his trip. Overindulgence stifles
growth by failing to provide the stimulus of frustration. This explains
more fully Kohut's position on reassurance or any other form of grati
fication by the analyst. Kohut viewed such behavior as contraindi
cated because it does not provide the optimal frustration necessary for
growth to occur. Only interpretation has the proper mixture of empa
thy and frustration to allow the transmuting internalization process
that was arrested in childhood to become completed.
The structure built from the transmuting internalization of the ide
alized analyst imago becomes the ideals of the patient. The patient
now has values and principles worth striving for because the idealiza
tion formerly invested in the analyst as an external figure is now
focused on the values and principles held by the patient. The narcis
sistic libido formerly invested in the external figure of the analyst is
now devoted to a part of the patient's own psyche. Once this process
is completed, the patient's ideals have been sufficiently strengthened
to function effectively as an integral pole of the self. The superego not
only exists but is held dear and becomes meaningful for the achieve
ment of the patient's life goals. If the patient had suffered from a more
primitive rupture in the effort to idealize the parent, the internaliza
tion of a solid set of ideals now helps to achieve a narcissistic balance
278 Chapter 6
from the inside, which heretofore was achievable only from the
attachment to an external figure. If the trauma was to the middle level
of idealization, the internalized set of ideals helps to create the neu
tralizing capacity that had been missing. Whatever the level of the
trauma in the early idealization of the parental figure, the transmuting
internalization of the idealized analyst results in a set of ideals and
values that are solid and meaningful and that function as an integral
part of the self.
Kohut illustrated his concept of the resolution of the idealizing
transference in his discussion of Mr. X. Like Mr. A, this patient pre
sented with a presumed sexual dysfunction. Mr. X masturbated with
homosexual fantasies but had had no sexual experience of any type
when he came to treatment at age 22. Like Mr. A, he was socially iso
lated and lonely. His mother was enmeshed with him and fostered a
depreciatory attitude toward the father. The patient wished to join the
ministry and had a clear identification with Jesus Christ. His mother
emphasized to him the story of the baby Jesus and his mother.
However, Mr. X could not join the ministry, as was his wish, because
he had sexualized his relationship with religion. He had homosexual
fantasies of his pastor, and his most powerful masturbatory fantasy
was of crossing his penis with that of the priest at the moment of
receiving Holy Communion:
At the moment of climactic ejaculation the patient's preoccupation with a

powerful man's penis, with oral incorporation, and with the acquisition
of idealized strength found an almost artistically perfect expression in
his sexualized imagery about the consummation of the most profoundly
significant symbolic act of the Christian ritual [Kohut, 1977, p. 201].
In the first phase of the analysis of Mr. X, the transference consisted

of the patient's grandiosity, as expressed in his arrogant behavior and
his identification with Jesus Christ. As the vertical split between his
grandiosity and his empty self broke down, it became clear that the
latter was more authentically his and the grandiosity an appendage of
his mother's self. After the vertical barrier was removed, the analysis
shifted to the second phase of making conscious "the unconscious
structures" that were defended by his lethargy and emptiness. The
longings that lay buried were depicted initially in the following day
dream: The patient is driving his car while the engine sputters and
eventually stops working completely. He realizes he has run out of
gas. He is unable to obtain help from passersby but remembers that he
stashed away a gallon of gasoline in his trunk under a pile of junk. He
Heinz Kohut 279
pours the gas into his tank as the daydream ends. It developed in the
analysis that on occasion Mr. X had taken long walks with his father
during which his father told him stories of his hunting skill and success.
At these times he greatly admired his father as a teacher and guide, but
the pair never spoke of these walks afterward, and they remained iso
lated in Mr. X's life. Kohut interpreted the gasoline fantasy as the
patient's expression of his need to be helped to find his nuclear self,
which was formed on the basis of his idealized relationship with his
father but long ago buried.
The second phase of the analysis consisted of making conscious
this idealization of the father and its manifestation in the idealizing
transference, which became the central transference theme of the
analysis. Once the idealizing transference became established, the
focus of the analysis became "(a) merger with the paternal ideal, (b)
de-idealization and transmuting internalization of the idealized
omnipotent selfobject, and (c) integration of the ideals with the other
constituents of the self and with the rest of the personality" (Kohut,
1977, p. 217). As these ideals became integrated into the self, the
structural defect was filled and the erotized enactments gave way.
The homosexual fantasies were a temporary means to attempt to fill
the defect in the self left by the burial of the idealized parental imago.
The analytic process replaced the need for the fantasies by activating
the idealization in the transference and working it through. In this
way, Mr. X reactivated the real childhood relationship with his father.
Kohut (1977) concluded as follows:
It was with the aid of the analytic work focused on the sector of his person
ality that harbored the need to complete the internalization of the idealized
father imago and to integrate the paternal ideal, after the analysis had
shifted away from preoccupation with Mr. X/s overt grandiosity, that struc
tures began to be built, that a firming of the formerly isolated, unconscious
self could take place through gradual transmuting internalization [p. 218].
Kohut viewed the case of Mr. X as a demonstration of the treatment

of a patient with narcissistic personality disorder and a predominantly
idealizing transference structured with the typical combination of
vertical and horizontal splits. It should be noted that Mr. X's grandios
ity was manifest long before the need for idealization but that the lat
ter, deeply repressed, was more crucial for the therapeutic process. It
was the transmuting internalization of the activated need for the ide
alized father that led to the formation of new psychological structure.
This fact also illustrates Kohut's (1977) principle that in narcissistic
280 Chapter 6
pathology the goal of treatment is to strengthen one pole of the self

enough that the self is able to function. Once Mr. X's idealization pole
was sufficiently strengthened, his homosexual preoccupations abated;
and at that point he had the psychological organization necessary to
function socially and harmoniously. Having illustrated the therapeu
tic process with regard to the idealizing transference, we now turn to
the process of structure formation resulting from the therapeutic
transformation of the mirror transference.
In Kohut's view, the manifestation of the grandiose self in treat
ment leads to a mirror transference. Kohut referred to both a mirror
transference in the broad sense and a mirror transference in the nar
row sense, although some of his followers have recommended that it
be given a single usage (Wolf, 1988). In Kohut's terminology, mirror
transference in the broad sense refers to the needs of the grandiose self
in relation to the analyst at any of its three levels. In its most primitive
form, the patient seeks a merger, as the boundaries of the grandiose
self extend to include the analyst; in the merger transference, such
patients expect to have the same degree of control over the analyst as
they do over the parts of their own body or mind. In the intermediate
form of mirror transference the analyst is expected to share the same
characteristics and attitudes as the patient; Kohut calls this form of
mirror transference the "twinship transference." The highest level is
the mirror transference in the narrow sense, or the mirror transference
proper, which includes the need of the grandiose self for the analyst to
confirm its value by giving approval and admiration.
In all three forms the analysis becomes focused on and organized
around the needs of the grandiose self to be confirmed and admired.
In an analogous manner to the idealizing transference, the analyst as
an external figure is providing the narcissistic function that patients
cannot provide for themselves. In this way, patients experience self
cohesion via their attachment to the analyst. The analysis therapeuti
cally activates the archaic childhood grandiosity, and the analyst sus
tains it by fulfilling the selfobject function not provided by the parents
in childhood.
Does this mean that the analyst should mirror the patient by pro
viding the approval and admiration that were not received in child
hood? Kohut's answer to this question is an unequivocal no. Any such
effort by the analyst would gratify the patient's childhood longing
rather than help resolve it. This is a common misunderstanding of
Kohut's approach; therapists frequently speak of "mirroring the
patient" as though the provision of admiration and approval were a
component of Kohut's treatment approach (Wolf, 1988). The analyst's
Heinz Kohut
role is to foster the emergence of the archaic grandiose self but not
to gratify it. Patients form a mirror transference, but the role of ana
lysts is no more to mirror than their role in the idealizing transfer
ence is to make themselves idealizable. Patients feel mirrored by the
analyst's empathic understanding, and to the extent that they do
not, their reactions are interpreted. It is the interpretation of frus
trated needs for mirroring that becomes the heart of the therapeutic
action. This process should not be disturbed by the provision of
artificial gratifications.
The same principle applies here as in the case of the idealizing
transference: the point of the therapeutic activation of the grandiose
self in the form of the mirror transference is to relinquish it gradually
and thereby transform the grandiose self via transmuting internaliza
tion into psychological structure. As is true of the idealizing transfer
ence, this transformation can only occur if the analyst provides
optimal frustration. In his early work, Kohut (1971) attributed optimal
frustration to the inevitable failures in the analyst's empathy. Later,
Kohut (1984) pointed out that even accurate interpretations, although
they provide understanding, are inherently optimally frustrating
because they do not gratify directly the unmet needs of childhood.
According to this later view, the patient always wants more than the
interpretation provides. Since the analyst's empathy with the needs of
the grandiose self is not tantamount to the meeting of those needs,
analytic empathy is the means for the transformation of archaic
grandiosity into new psychological structure.
Kohut's view is that the mirror transference, like the idealizing
transference, will spontaneously establish itself if given the opportu
nity by the analyst. The analyst's role early in treatment is not to inter
pret the patient's grandiose-exhibitionistic needs and is certainly not
to interfere with them in any way. Any educational measure designed
to tame the patient's grandiosity by labeling it "entitlement" or any
other pejorative will injure the fragile grandiose self and may lead to
its repression. The analyst's role is to accept the mirror transference
and the patient's tendency to use him or her as a selfobject as the
expression of a developmentally appropriate need.
The therapeutic mobilization of the mirror transference is repre
sented by the analysis of Miss F, whom Kohut (1968) treated when he
was still operating within the traditional analytic framework; indeed,
some have given this case the pivotal role in changing Kohut's theo
retical orientation from ego psychology to self psychology (Ornstein,
1978). During prolonged phases of the analysis of Miss F, sessions
developed a repetitive pattern. In the initial part of the session the
282 Chapter 6
analysis seemed to be progressing smoothly: the patient appeared to

free associate easily and to be engaging in self-analysis. However,
Kohut (1968) eventually made a significant observation:
I came to the crucial recognition that the patient demanded a specific

response to her communications and that she completely rejected any
other. Unlike the analysand during periods of genuine self-analysis, Miss
F could not tolerate the analyst's silence; at approximately the midpoint
of the sessions, she suddenly became angry at me for being silent [p. 504].
When Kohut summarized what she said Miss F felt calm, but if he
went even slightly beyond her comments she became enraged and
accused Kohut of ruining the analysis. Kohut eventually came to view
this pattern as the emergence of the patient's grandiose self and its
concomitant need for mirroring. In this mirror transference the patient
did not view the analyst as a separate person but as a functional part
of her grandiose self. In the patient's view, the analyst's role was to
admire and echo her.
According to Kohut, the treatment took the proper course when he
realized that this mirror transference was Miss F's attempt to integrate
her archaic grandiosity with the rest of her personality. Miss F's pre
sumed self-analysis masked her need to be echoed and admired. In
Kohut's view, the effort to use the analyst in this way is not a "resis
tance" but the spontaneous mobilization of a phase-appropriate devel
opmental need. The patient's "narcissistic demands" were her effort
to complete the arrested development of her self. The use of the ana
lyst to mirror her was her effort to achieve in the analysis what was
missing from her childhood: a functional reliance on a selfobject to
strengthen the developing self. For the crucial issues to be mobilized
in the transference, the analyst had only to avoid interfering with their
spontaneous appearance.
The case of Miss F also illustrates Kohut's approach to defense and
resistance, which will be discussed further later. As can be seen from this
case, Kohut's belief that narcissistic pathology consists of arrested self
development leads to an appreciation of the patient's need to form the
selfobject transference rather than to receive interpretations of resistance.
Miss F was not "resisting," in Kohut's view, but attempting to get her
thwarted childhood needs met. The patient's demands represent a posi
tive move and must be interpreted as such by the analyst for the full
mobilization of the transference. To have interpreted Miss F's demands
as resistance would have been a technical error and would have risked
the resubmergence of her efforts to resume self-development.
Heinz Kohut 283
In Kohut's view, there are two additional advantages to the ana

lyst's acceptance of the phase appropriateness of the mirror transfer
ence: it eliminates iatrogenic anger, which tends to result from a
critical therapeutic attitude, and it stimulates childhood memories of
thwarted narcissistic needs. Both advantages are demonstrated in the
case of Mr. Z (Kohut, 1979). This is Kohut's only published case of two
complete analyses of the same patient, one from the traditional per
spective and the other from the viewpoint of self psychology. Mr. Z
was socially isolated—had no girlfriends, he had only a single male
friend—and lived with his widowed mother. He sought analysis after
his only friend found a girlfriend and distanced himself from him. The
patient's only sexual involvement was masochistic masturbatory fan
tasies in which he was a slave to domineering women. In the first
analysis Mr. Z demanded that the analyst be under his "exclusive con
trol"; Kohut interpreted that he wanted to be "admired and catered to
by a doting mother who . . . devoted her total attention to the patient"
(p. 5). The patient responded to this interpretation with explosive
rage. Indeed, the first one and one-half years of the analysis were con
sumed with Mr. Z's rage at Kohut, which subsided only after Kohut
made the comment, "Of course it hurts when one is not given what
one assumes to be one's due" (p. 5). After the patient calmed down,
the analysis entered the second phase, which was dominated by the
Oedipus complex, as Kohut then understood it. The analysis centered
upon Mr. Z's castration anxiety, his masturbatory fantasies, his fan
tasies of the phallic woman, his witnessing of the primal scene, and a
homosexual relationship with an adult man when the patient was 11
years old.
Although Kohut felt the analysis had been a success, the patient
contacted him five years later because his relationships were still
emotionally shallow; he received no joy from sex; and he did not
enjoy work, although he was successful. In the second analysis the
patient once again demanded "perfect empathy," became enraged at
the slightest misunderstandings, and adopted an attitude similar to
the first phase of the first analysis. The difference was Kohut's (1979)
attitude: "I looked upon it as an analytically valuable replica of a
childhood condition that was being revived in the analysis" (p. 12).
This time there were no complaints of being misunderstood and no
rage reaction from the patient. Instead, there appeared a stream of
childhood memories that had been absent from the first analysis.
"This phase of the analysis revived the conditions of the period
when, in early childhood, he had been alone with his mother, who
was ready to provide him with the bliss of narcissistic fulfillment at
284 Chapter 6
all times" (p. 12). Kohut now viewed the patient's demands as his
struggle to disentangle himself from the mother, a noxious selfobject
who had later become psychotic.
As a result of these memories, a whole new sector of the patient's
personality—the arrested development of the self—opened itself for
analysis. The patient's traumatic disappointment in his father; his life
long yearning for an idealized, strong replacement; his need to break
away from the merger with his delusional mother—these were the
elements that occupied the center of the second analysis. Kohut
viewed the dramatic shift in the analytic material as the result of his
acceptance of the phase appropriateness of the patient's narcissistic
demands and his willingness to take them seriously rather than
attempt to get rid of them by interpretation. He concluded that his
interpretive approach toward the patient's narcissistic demands in the
first analysis had conveyed a moralistic, disapproving stance and had
led to Mr. Z's compliance with his theory, just as the patient had com
plied earlier with his mother's delusions. When Kohut shifted toward
understanding the patient's demands, Mr. Z was able to open his
analysis to his childhood longings for the idealized father and to his
merger with the delusional mother, from which he could now dare to
free himself.
Kohut recommended that grandiosity not be interpreted when its
functioning is smooth because the analyst is serving as the mirror
ing selfobject needed to achieve self-cohesion. However, just as with
the idealizing transference, this function will inevitably be dis
rupted by major and minor "failures" in analytic empathy. In anal
ogous fashion to the analysis of an idealizing transference, the
therapeutic action begins when the selfobject fails. As discussed ear
lier, such empathic failure is most commonly seen in seemingly
minor events when the analyst misunderstands a communication or
fails in any way to be perfectly in tune with the patient. Major fail
ures include vacations or other interruptions of the treatment. Any
of these events will be experienced as an injury and will disturb
the patient's narcissistic equilibrium. The trauma here is loss of the
mirroring function on which the cohesiveness of the self depends.
The threat to the self gives rise to disintegration anxiety, and the
patient's response to the empathic failure will be either shame or
rage followed by a regressive reaction. Patients who feel shamed by
the revelation of a defect may become depleted and show signs of
depression and lack of energy or may become preoccuped with
bodily tensions, resulting in hypochondriacal symptoms. Or the
patient may become enraged at the "offense," resulting in intense
Heinz Kohut
and fragmented rage and often a need for vengeance (Kohut, 1972).
A good example of a regressive reaction to the analyst's vacation by
a patient with a mirror transference is the case of Mr. K (Kohut, 1971).
After a brief period of idealization, this patient formed a merger/
twinship transference in which the importance of the analyst lay in his
function as an extension of the patient, and at times in being like the
patient. Whenever the analyst went on vacation, Mr. K felt empty,
withdrawn, and depressed. In addition, his dreams shifted strikingly
from people imagery to machines, electrical wiring, and, often, spin
ning wheels. Mr. K was not aware of having an affective reaction to
the impending separation from the analyst. However, it developed in
the analysis that the machine imagery and especially the spinning
wheels represented parts of the body and that the impending separa
tions had given rise to a regression to hypochondriacal preoccu
pations. This regressive reaction was analogous to the patient's
investment in his body when as a child he was faced with narcissistic
trauma. For example, at age three, his brother was born and his
mother, unable to cope with more than one child at a time, abandoned
the patient emotionally. He turned to his father, but the mother's inter
ference and the father's inability to tolerate idealization made this solu
tion impossible. The boy attempted to discharge his narcissistic tension
by engaging in physical activities. When the analyst announced a vaca
tion, Mr. K regressed to bodily preoccupations in order to discharge
the narcissistic tension created by the impending absence of the selfob
ject who served the mirroring function. It should be emphasized that
Mr. K did not miss the analyst as a person. Rather, he felt a loss of a
self function, a part of his self, a loss that led to narcissistic tension. Mr.
K attempted to resolve the tension through bodily means, and this was
represented in the dreams by machine imagery.
Whether the patient becomes ashamed or enraged, the analyst's
task is to discover the source of the injury and interpret it. Just as with
the idealizing transference, the active interpretative work begins with
the disruptions to the patient's narcissistic balance due to disturbances
in the selfobject transference. The analyst's role is both to uncover the
source of injury and interpret its phase appropriateness. That is, the
historical source of the narcissistic deficit must be shown to patients
so that they can see that the analyst accepts their reaction as an under
standable response to an unmet childhood need. The dual function of
discovering the source of injury and understanding its developmental
context is the essence of the explanatory component of interpretation.
Kohut's understanding of the meaning of the mobilization of grandi
osity in the treatment situation is in clear opposition to Kernberg's
Chapter 6
treatment approach. For Kernberg, as we saw in chapter 5, the gran

diose self is a pathological, defensive construction, whereas for
Kohut it is a normal, phase-appropriate developmental step that has
become arrested due to trauma. Therefore, while Kernberg recom
mends immediate interpretation of grandiosity as a defense, Kohut's
therapeutic approach is to allow the full development of the gran
diose self and to appreciate the appropriateness of its reemergence
in view of the developmental context in which the patient was trau
matized. In Kohut's view, this technical approach provides the
empathy that was not received in childhood and that is now needed
in adulthood to resolve the childhood trauma. From this perspective,
Kernberg's approach of interpreting the grandiose self as a defense
would serve only to bury the patient's frustrated longings, which
need to be therapeutically activated to be resolved.
The empathic response of the analyst allows the mobilization of the
patient's grandiose self within the mirror transference. The treatment
process becomes a continual cycle of mirror transference, disruption,
interpretation, and restoration of the mirror transference. Each time
the cycle occurs, a bit of the grandiose self becomes transformed into a
structure of realistic self via transmuting internalization. In this way,
the nuclear self becomes invested with narcissistic libido. If the ana
lyst continues to respond empathically, the patient is gradually able to
transform the grandiose self into realistic ambitions and a healthy
sense of self-esteem, both of which are now invested with the positive
self-enhancing feeling that initially inhered in the grandiose self. As
ambitions become realistic and solid, new psychic structure is formed
and one pole of the self is strengthened.
Despite his copious clinical illustrations, Kohut did not provide
detailed descriptions of the working-through process of the mirror
transference, that is, the process of transformation of archaic grandios
ity into realistic ambition. The best such description is probably the
case of Mr. I, a patient who was treated by one of Kohut's colleagues
(Goldberg, 1978). This young man sought analysis because he felt that
he was too preoccupied with his unhappiness and the lack of direction
in his life to perform well on his job. He had difficulty reading or sit
ting still long enough to complete his work projects. It became clear
that he had a frenetic lifestyle that included addictive masturbation,
preoccupation with one-night sexual exploits, and sitting on the toilet
frequently and for prolonged periods. He had to date frequently but
was unable to form and maintain a relationship.
Mr. I immediately settled down in the analysis, and his sexual
obsessions subsided. He became quickly connected to the analyst,
287
toward whom he felt like a "splat on a wall." He became preoccu

pied in the analysis with his desire that the analyst be excited by
him, and he wished the analyst to w itness his every act. He
expressed the wish for extra-long sessions and once verbalized the
wish that the analyst would take one month just to listen to him.
Mr. I took all this as evidence of his "addiction" to the analysis. He
felt that the analyst was a powerful figure who provided him with a
stable, safe anchor, and he began to feel that the analyst was "per
fectly in tune" with him. The analyst's response to him became the
most significant issue in his life, even though he hated his depen
dence on the analyst. Before the analyst's first vacation Mr. I had a
strong reaction and developed homosexual fears. During the vaca
tion Mr. I assumed a gross identification with the analyst, treating
his girlfriend with the kindness, understanding, and tact he himself
felt from the analyst, but had not heretofore shown to her. The
patient was quite excited that he and the analyst "hit it off" after the
vacation, and at that point his masturbation, desperate need for
dates, and prolonged toilet sitting all markedly declined.
The first phase of Mr. l's analysis, a period during which he mar
ried, focused on the idealizing transference, but our focus in this dis
cussion is on the process of resolution of the mirror transference. After
the idealizing transference had been worked through, Mr. I began to
discuss the possibility of termination, although this was partly forced
by an impending job change. The anticipation of termination led to
the establishment of a mirror transference as Mr. I developed an inten
sification of his sensitivity to separation and to the analyst's response
to him. He became consumed by the topic of separations; weekends,
vacations, and termination were the primary issues in the analytic ses
sions. During a Christmas vacation, Mr. I felt "disconnected," drab,
and colorless. The analyst commented on the positive aspect of this
response and how it meant that he was able to keep himself calm
without the analyst, although he eventually "overshot the mark and
then began to feel the emptiness and monotony" (Goldberg, 1978,
p. 82). The patient was so moved by this remark that he felt like cry
ing, and the analyst commented that he was overstimulated by the
experience of feeling understood after not having been understood for
so long. The patient was also struck by the truth of this remark and
felt moved once again. The impact of this understanding was endur
ing. Mr. I had the impulse to tell the analyst to "cut it out" but later
realized that he could now react without falling apart.
Over the next weekend Mr. I dreamed that he put gas in his car but
could not shut it off and that it spilled over. The dream was interpreted
288 Chapter 6
as the patient's sensitivity to the analyst's interventions. The patient

responded by verbalizing his fear that termination meant flying off
into space and that his "equipment" would not function and he would
crash. The analyst pointed out that whenever Mr. I felt his landing
instruments or the analyst's radar were defective, he feared "losing
contact with ground control."
The next day the patient reported "a dramatic change," which indi
cated to him that the analyst's sensitivity must have been just right.
He also had an elated response to a talk by a senior executive at his
firm, feeling that the implication was that "father and son should
work together." He felt that the previous day's session had indicated
that the analyst's "radar" and "ground control" were functioning well
and that he was not going to crash land. Mr. I went on to say, "[I
need] a base of operations, a relationship to give meaning to all my
activities. I don't need as much as I used to from somebody else. But if
I don't get it I feel let down, sick, and futile" (p. 85). The analyst did
no more than agree, and when over the next few sessions the patient
was frequently silent, the analyst was also. Mr. I felt the silences were
accomplishments, as it had taken him three years of analysis to be able
to be silent.
This process seemed to result in a dramatic shift in Mr. I. He told
the analyst, "[I will] take this unit, you and me, put it in a capsule,
and bury it somewhere deep inside me—it will always be there and
give me self-confidence and consistency" (p. 87). By this comment,
the patient seemed to be concretely representing his ability to take
in the analyst and his mirroring function and make them a part of
his own psychic structure. Soon after, he began to discuss a specific
termination date. He then reported that he no longer needed to sit
on the toilet for long periods of time. He realized that the prolonged
toilet sitting had been one more way of getting pleasure and excite
ment and that he no longer felt the need for this. When he had a
birthday during this period, he described feeling a lack of "fanfare"
no "bands playing" and reminisced about the difficulty in giving up
the desires of the past and putting the analysis into the past. When
the analyst next went on vacation, the patient had a good experi
ence which was represented in a dream that he hugged a boy who
had epilepsy and who suffered no seizures if hugged. The interpre
tation was that the patient was both the boy and himself in the
dream, that is, that he could now soothe himself. At this time Mr.
I and his wife gave a party, and he felt no need for special atten
tion; this incident made him recall how desperately he had needed
attention at the beginning of the analysis.
Heinz Kohut 289
When Mr. I did set a definite termination date for two months
after the party, he became depressed and scared that the ending
would be "too much." Although he felt that there were no further
content issues left in his analysis, he was frightened that without
the analysis his stability was at risk. The patient became highly
sensitive once again to the analyst's interventions. He told the ana
lyst that he needed him to listen again and not say too much. He
felt that he could now say anything and felt more "equal" to the
analyst. Toward the end of the analysis, Mr. I dreamed of a man
who swallowed a clarinet that played from inside him. He felt that
he was taking the analyst with him, whole. The analyst pointed out
that he would have to digest the analyst and that the music came
from inside him. In the last few weeks of the analysis, Mr. I
reviewed his accomplishments in the treatment. He believed that
the "underlying force" was the stability of the analytic relationship
and that this had allowed him to transfer his ties to others. He felt
that he was different "every minute of the day" and that he now
could do for himself what the analyst had done for him.
This case illustrates the working-through process of the mirror
transference according to the principles of self psychology. The
patient's functioning depended on the stability of the analytic rela
tionship, specifically, the mirror transference. All the disruptive
analytic experiences were part of the transmuting internalization
process, which changed the narcissistic need, in this case to be mir
rored, into psychological structure in the form of realistic ambition.
Early on in the analysis, the analyst replaced the masturbation and
sexual addictions. By the end of the analysis, Mr. I no longer
needed the analyst to soothe him, reduce his tension, or perform
any other function. The transmuting internalization process had
been completed.
The therapeutic action in the analysis of the grandiose self, as in
that of the idealized parental imago, is a combination of acceptance
and even welcoming of the archaic, repressed, or split-off self and
the interpretation of the narcissistic injuries it receives once it has
become activated in the transference. In this way, the analytic
process allows the arrested self to resume development by trans
forming archaic grandiosity into mature self structure with realistic
self-esteem. Does this view of the analytic process apply only to
narcissistic disorders or is it to be taken as a new model for psycho
analytic treatment in general? To address this question, we must
understand Kohut's view of the self-psychological treatment of the
structural neuroses.
290 Chapter 6
Treatment of the Structural Neuroses
Kohut's view of neurosis as a disorder of the self leads us to consider

the implications of self psychology for the treatment of neurotic condi
tions. First, self psychology reformulates as narcissistic disorders
those cases that are viewed as oedipal neuroses in the classical model.
We have seen that Kohut believed that the drive-ego model led ana
lysts to view oedipal conflicts and drives as psychological bedrock
when, in fact, deeper anxieties and issues frequently lay beneath them.
Although cases with clear oedipal issues tend to be diagnosed as neu
rotic disorders just as in the classical model, the emphasis of self psy
chology on threats to self cohesion and on disintegration anxiety
implies a tendency to see oedipal material as a manifestation of a
defective self. This means that Kohut diagnosed many cases as narcis
sistic disorders that appeared to be oedipal neuroses and focused the
treatment on the resolution of the selfobject transference rather than
on an oedipal drive-dominated transference.
The case of Mr. Z illustrates this implication of self psychology.
Recall that in Mr. Z's first analysis he was treated as a neurotic case,
his demands were treated as resistances, and his treatment focused on
his fantasies of the phallic woman, castration anxiety, masturbatory
fantasies, and his witnessing of the primal scene. His homosexual pre
occupations were interpreted as a regression from the oedipal compe
tition with his father and the resulting castration anxiety. In the
second analysis the same symptoms were interpreted as products of
the patient's inability to feel alive. His stimulating masturbatory fan
tasies counteracted his feeling of deadness and replicated his child
hood masturbation, which he used to soothe his joyless, depressed
childhood. In the second analysis Kohut understood Mr. Z's homosex
ual relationship with the older man in childhood and his adult homo
sexual interest as his yearning for a relationship with an idealized
father figure. Kohut concluded that competition with the father and
castration anxiety had no appreciable causative role in the homosex
ual masturbatory fantasies and interest. He interpreted all the oedipal
material, including fantasies of the phallic woman and castration anxi
ety, as products of the patient's desperate effort to escape internal
"deadness" and feel alive by stimulating himself.
In the second analysis Mr. Z's demands for "perfect empathy" were
viewed not as a resistance but as a desperate effort to fill a gap in
his self structure, that is, as evidence of a narcissistic deficit. Conse
quently, his longings and demands were accepted and understood as
developmentally phase appropriate. The result was a spate of material
Heinz Kohut 291
about the patient's mother's domination of him, including her preoc

cupation with his feces, possessions, and skin blemishes. This material
had not appeared in the first analysis; it developed in accordance with
the patient's longing for an idealized figure, which he found in the
analyst. In the second analysis, the oedipal issues and masturbatory
fantasies of strong women were not prominent analytic material and
became resolved once the self developed a more stable structure. This
is typical of the way Kohut regarded oedipal conflicts and what classi
cal analysts would call drive-related issues. He tended to focus on the
selfobject transference and felt the drive issues would abate once the
self was strengthened.
As we have seen, Kohut rejected the view that all structural neu
roses may be reformulated as narcissistic disorders and maintained a
belief in the value of the diagnosis of structural neurosis in certain
cases, presumably fewer than would be so designated by classical ana
lysts. This brings us to the clinical implications of self psychology for
cases that fit the diagnosis of structural neurosis.
Because Kohut (1984) viewed structural neuroses as products of
arrested development due to faulty self-selfobject relationships in the
oedipal phase, he felt that it was necessary for the analytic situation to
engage the thwarted needs of this phase. Just as narcissistic patients
must bring forth their archaic narcissistic configurations, so too must
neurotic patients mobilize the needs with which they entered the oedi
pal phase. If the treatment situation is to mobilize the frustrated needs
of the oedipal phase, the joyful assertiveness and affection of the
child's entry into this phase, together with the expectation of confir
mation and approval, must be mobilized in the transference. Again,
for Kohut, it is not the Oedipus complex with its lust and hostility that
needs to be enacted in the analysis but the selfobject needs associated
with the child's entry into the oedipal phase.
Kohut believed that the analysis must retrace the steps of the neu
rosis to arrive at the thwarted developmental needs. Initially, there is
a period of severe resistance to the Oedipus complex and castration
anxiety. Once these resistances are worked through, the Oedipus com
plex with its hate, death wishes, and lustful longings appears. In
Kohut's view, traditional analysis stops at this point and misses the
crucial issues of the oedipally damaged neurotic patient, since the
analyst has a preconceived notion that the Oedipus comlex is psycho
logical bedrock. To Kohut, the appearance of the Oedipus complex
masks resistances under which lies disintegration anxiety, the fear of
the crumbling of the self, which now becomes apparent. Under the
next mild level of anxiety lies a healthy, joyfully undertaken attempt
292 Chapter 6
to enter into the oedipal phase. That is, the analysis traces the devel
opmental steps of the neurosis in reverse direction, beginning with the
isolated drive fragments of the pathogenic Oedipus complex, then
uncovering the failure of the oedipal phase selfobjects, and finally
reaching the child's joyful anticipation of an assertive relationship
with the parent of the opposite sex and an affectionate one with the
same-sex parent.
With the mobilization of the child's joyful anticipatory entry into
the oedipal phase, the crucial transference constellation has entered
the analysis. The analyst's task is to understand that the patient's wish
for affection and self-assertion are healthy, phase-appropriate child
hood longings so that they do not become reburied under intense
resistances. In Kohut's view, the traditional model mistakenly views
such healthy wishes as defenses against lust and hostility, thereby
mistaking the patient's struggling effort to complete the development
of the self as a resistance to be overcome. By understanding and wel
coming the patient's effort to unblock the development of the self, the
analyst can help in the achievement of this goal. If these efforts are
mistaken for resistances, they are in danger of becoming repressed
once again and the development of the self remains incomplete.
Kohut provided no case descriptions of a self-psychologically
informed analysis of a neurosis, but one can discern his concept of
the treatment in his discussion of female agoraphobia mentioned
earlier. Recall that Kohut stated that the failure of the paternal self
object to welcome the little girl's affectionate and assertive strivings
in the oedipal phase resulted in their conversion to lust and hostility.
In addition, the failure of the maternal selfobject to provide soothing
functions led to a gap in the child's ability to soothe herself and, con
sequently, to a tendency toward disintegration anxiety and an easily
threatened self. The result was that the patient could not leave home
without another woman to provide the soothing she was unable
to provide for herself. It is clear from this formulation that Kohut's
treatment approach would be to allow the development of the
expected idealizing transference. As the transference becomes mobi
lized, the analyst would function to fill the gap in the self. The addic
tion to the woman should lessen as dependence on the analyst takes
its place. The analytic process would then involve the patient's reac
tions to disappointments in the analyst and their interpretation. The
transmuting internalization process, the means of resolution of all
such transferences, would be expected to result in the internalization
of self-soothing functions and the subsequent building up of self
structure.
Heinz Kohut
One can see that Kohut's views culminated in the belief that selfob
ject needs typify both the neuroses and the normal oedipal phase. This
view raises the question of whether selfobject needs are ever totally
overcome. Kohut (1984) maintained that selfobject needs continue
throughout life. This aspect of his thought is critical because it reflects
a shift in the goals of analysis away from the "independence morality"
he believed the traditional theory had imposed on it. Kohut believed
that the drive-ego model assumes that dependence is a negative trait
and that the goal of analysis is to free patients from attachments and
enable them to learn to achieve self-esteem without selfobject needs.
Kohut believed that such a view of the human process is an illusion
because even the psychologically healthiest individuals require posi
tive responses from others to feel a strong sense of self. He came to
believe that selfobjects are the oxygen of human life. Thus, the goal of
analysis is not to free patients from all need for selfobject contact but
to replace their "archaic" selfobject needs with "mature" selfobject
needs and enable patients to evoke the needed responses from the
mature selfobjects.
Kohut's extension of his treatment model to the neuroses raises the
question of whether the self psychological model of treatment ought
to replace the drive-ego model in the analysis of these milder disor
ders. While Kohut (1984) was reluctant to conclude that traditional
analysis of neurotic patients should be replaced by self psychology, he
increasingly believed in the applicability of the self-psychological
model of cure even for these patients. The curative process for the
neurotic patient involves selfobject transference and transmuting
internalization, just as it does for the narcissistic patient. The differ
ence lies in the developmental phase of the selfobject failure, which
has no significant impact on the treatment process. Curtis (1985), in
his criticism of self psychology, pointed out that although Kohut con
tended that the replacement of the classical model by self psychology
was a question for the future, his reformulation of the treatment
process accomplished this goal in the present.
The fact is that Kohut appeared to believe that the Oedipus com
plex could never be the cause of psychopathology of any type,
since the intense affects of lust and hostility become issues only if
the self is weakened, and that even the pathogenic sexual and
aggressive conflicts in cases of neurosis are products of the break
down of the self. Kohut did not believe that expression of a drive in
and of itself resolves conflict. According to Kohut, chronic rage and
sexual promiscuity are symptoms of a weak, enfeebled self that
yearns for selfobject responses and will continue such behaviors
294 Chapter 6
until it gets them. Once the self becomes strengthened by the

responsiveness of a selfobject, the "drive" needs are relieved.
It is clear from this conception of the psychoanalytic treatment of
the neuroses that Kohut did not believe that making the unconscious
conscious was the essence of psychoanalytic cure even for the struc
tural neuroses. Indeed, he criticized this traditional concept of the
analytic process as the imposition of the analyst's "truth morality"
on the patient. Kohut advocated a shift in analytic priorities from
the moralizing value system of the traditional analyst to the smooth
functioning of the self. Making conscious the unconscious may or
may not be necessary for the achievement of this goal in a given case.
When the unconscious is made conscious, according to Kohut, it
helps the neurotic patient, like the narcissistic patient, not because of
the increased awareness to which it gives rise but because a selfobject
has provided an experience of optimal frustration that culminates in
the development of new self structures.
This view raises the question of why traditional analyses have had
some success given the tendency of the traditional model to see
healthy longings as resistances. Kohut's answer is that any model
within the psychoanalytic tradition is successful to the extent that it
provides understanding and explanation. According to Kohut, the
classical analyst often conveys correct understanding and proceeds to
offer incorrect explanations. Kohut (1984) gave as an example an inter
pretation by a Kleinian analyst, but the principle applies to traditional
analysis as well. In the example, the patient's withdrawal after a can
celled session was interpreted as transference anger due to his view of
the analyst as a bad breast. Kohut believed the interpretation to be
incorrect but was surprised to learn that the patient felt much better in
response to it. He explained this apparent anomaly by the fact that the
analyst correctly understood the patient's withdrawal as a response to
the canceled session; that is, empathic understanding provided relief
despite the erroneous explanation invoking the bad breast. Thus,
many contradictory theoretical schools can promote beneficial under
standing. However, complete analytic cure requires both empathic
understanding and accurate explanation.
Although Kohut was reluctant to state definitively that classical
analysis should be abandoned in favor of a self-psychological
approach in the treatment of neurosis, it is clear that he gave the
drive-ego model no primary role in treatment. By approaching the
analysis of neurotic conditions as the restoration of the self's normal
developmental process, Kohut had, in essence, replaced the classical
model with self psychology as both an explanatory system and a treat
Heinz Kohut 295
ment method. By the time of his death in 1981, Kohut's views had
evolved into a new psychoanalytic model encompassing development,
psychopathology, and the treatment process.
SUMMARY: THE SELF PSYCHOLOGY MODEL
Ultimately, Kohut's thought led him to the creation of self psychology

as a new model within the analytic framework. Freud (1915a) had
developed psychoanalysis on the basis of the fundamental concept of
the drive, a border concept between the biological and the psychologi
cal. For Kohut, not only can drives not be the basis for a psychology of
complex mental states, but they cannot even be a part of such a psy
chology because they cannot be known by empathy and introspection.
Psychological life begins, according to Kohut, with the empathy of the
selfobject responding to the infant. The infant is born with innate
potentials, and when these givens come into contact with the empathy
of the first selfobject, a rudimentary nuclear self is born. Psychological
development is from the beginning concerned primarily with the self,
and the crucial factor in each phase is its relationship with its selfob
jects. Optimally frustrating selfobjects give the self strength, harmony,
and cohesion. Any faulty selfobject experience weakens the self and
leaves it prone to pathology. With these concepts, Kohut reoriented
psychoanalysis by directing its focus on the self, or, more precisely,
the self-selfobject relationship.
Since the fundamental issue in psychological development is the
growth of the self, pathology of whatever type is due to a develop
mental arrest that leaves the self weak, vulnerable, and prone to frag
mentation. Defenses are often erected to protect a vulnerable self, but
the essence of pathology does not reside in the use of defenses. In fact,
defenses should be appreciated by the analyst as self-preservative
maneuvers. With these concepts Kohut, echoing Winnicott's concept
of pathology as blocked maturational process, replaced the drive-
defense model with the pathology of developmental arrest. Since, in
his view, the self depends on its relationship with selfobjects and since
its defects are caused by disturbances in that relationship, all pathol
ogy ultimately derives from faulty selfobject empathy in crucial
phases of development.
This reconceptualization of pathology fundamentally changes the
nature of psychoanalytic treatment. Since pathology is arrested self
development, the crux of analysis is not interpretation of defense but
mobilization of the arrested self via empathic engagement and the
resumption of self development by transmuting internalization. The
296 Chapter 6
analyst acts as a selfobject who functions as a part of the arrested self

until he or she is internalized to complete its development. Inter
pretation is still the crucial tool of the analyst not so much because it
makes the unconscious conscious but because it provides the optimal
frustration necessary for the resumption of self development by trans
muting internalization. Most crucially, to help the self complete the
unfinished tasks of development the analyst functions in a manner
analogous to how the parental selfobjects should have functioned. To
the degree that the analyst allows himself or herself to be used to finish
the developmental task, the analysis will be successful.
The fundamental attitude of the self-psychological analyst is mark
edly different from that of the classical analyst. The latter sets himself or
herself in alliance with a part of the patient, the presumably healthy or
observing part of the ego, in order to break through the patient's resis
tances by interpretation and thus reach the unconscious wishes that lie
buried beneath the defenses. This concept has been a crucial component
of the psychoanalytic model of treatment since Freud (1895) first began
to trace associations back to the "pathogenic nucleus" of the neurosis.
In Kohut's model there are no resistances, and so there is no alliance
with one part of the patient's personality against another part. The self-
psychological analyst does not oppose the patient, not even a part of the
patient. Rather, the analyst's task is to be empathic with whatever mate
rial the patient brings to the treatment. If the patient is defensive, the
analyst's role is not to interpret the defenses in order to get rid of them
but to function as an empathic selfobject so that the defenses eventually
become unnecessary.
Kohut's analytic posture is the decisive feature that sets his views
into unalterable opposition to the classical model. The use of interpre
tation as a means for the achievement of transmuting internalization
and the appreciation of defenses as phase appropriate combine to
change the essence of analysis from the investigation of the uncon
scious to the completion of the self. Kohut, like Winnicott, believed
that analysts must allow themselves to be used according to the
patient's developmental needs of the moment. The essence of psycho
analysis, in this view, is not investigation but the development of the
self, and the role of the analyst is not to convey knowledge but to be
used as a selfobject. To the extent that the analytic relationship
becomes problematic, the analyst has failed in this function, and it is
incumbent upon the analyst to search for the empathic lapse rather
than to interpret the deficit that caused the patient to need the analyst
to be an unfailing selfobject. This concept of the psychoanalytic
process has become the focal point for much of the expansion of self
Heinz Kohut 297
psychology by Kohut's followers, but before we discuss the growth of

self psychology after Kohut, a critical assessment of some of his major
views will be offered.
CRITIQUE
Kohut made an enduring contribution to psychoanalytic theory by

introducing the concepts of the self and self-esteem in ways that have
significant therapeutic value. He contributed to the movement to
make psychoanalytic theory closer to patient experience with his dis
tinction between "experience near" and "experience distant" theoriz
ing and his application of it to his theory of self pathology. Kohut also
introduced into psychoanalytic theory a far more detailed account of
the development of the self and its vicissitudes than did any other the
orist, including those who, like Guntrip, employed the concept of the
self. Kohut also brought into psychoanalytic theory recognition of the
need for relationships with others in the struggle to achieve self
esteem and of the importance of these relationships throughout life. A
seminal clinical contribution is his recognition that much transference
enactment, which in the classical model is considered resistance, is in
fact the patient's expression of legitimate needs. Thus, Kohut fur
thered the movement toward appreciating the adaptive value of
patients' defenses and their survival value in the treatment process.
Kohut's ability to extend this thinking to an alternative theory of
development, pathology, and treatment has led to the most significant
and comprehensive new model of psychoanalysis since Klein's formu
lation. In two ways Kohut's alternative model is more radical than
Klein's: (1) self psychology attempts to establish psychoanalysis as a
pure psychology, that is, without the assumption of biological drives,
and (2) it proposes an alternative to the defense/resistance model. In
this sense, Kohut's model is a revolutionary movement within the psy
choanalytic tradition. Kohut's approach is a coherent, elegant alterna
tive that fits the child development data much better than the classical
model (see, e.g., Lichtenberg, 1983). Nonetheless, Kohut's model does
have deficiencies that must be addressed if it is to endure as an alterna
tive model or become the preeminent model within psychoanalysis. It
has been widely criticized by more traditional analysts as being "unan-
alytic" (Kernberg, 1982; London, 1985; Curtis, 1985, 1986; Rubovitz-
Seitz, 1988). Kohut's emphasis on the self rather than drives, on
empathy rather than making the unconscious conscious, and on
arrested development rather than defense interpretation have all been
attacked as fostering defense and resistance rather than working them
298
through. As indicated, Kernberg has been sharply critical of Kohut's

work. The emphasis in our critique of Kohut's model will be on three
major difficulties: confusion regarding the concept of the selfobject,
the homogeneity of psychopathology, and the lack of clarity in the
treatment approach.
Initially selfobject had a very specific meaning: an object used as a
part of the self. In this sense, the selfobject is similar to Winnicott's
notion of a transitional object (see chapter 4). In fact, when he was
still focused exclusively on pathological narcissism, Kohut (1971, p.
33) referred to the selfobject as a transitional object. This definition of
the term conflicts with Kohut's frequent tendency to equate selfob
jects with people. Kohut's work is replete with references to patho
genic failures of early selfobjects. Sensitive to the criticism that he
was leading psychology away from the psyche into interpersonal
relationships, Kohut (1984) made a point of insisting that the selfob
ject was an intrapsychic concept, not an interpersonal one. One can,
therefore, discern three definitions of the term: a transitional object, a
person in certain roles, and an intrapsychic concept.
Wolf (1988) attempted to defend Kohut against the charge of "inter
personalism" by contending that Kohut always insisted that selfobject
was an intrapsychic term. Kohut used the term to refer to parents and
analysts in their behavior toward the child or patient. For example,
Kohut (1984) attributed agoraphobia to "faultily responsive" paternal
and maternal selfobjects and then went on to say, "The mother, in
other words, was apparently not able to provide a calming selfobject
milieu for the little girl, which via optimal failure, would have been
transmuted into self-soothing structures capable of preventing the
spread of anxiety" (p. 30). In this formulation of agoraphobia, Kohut
used the term selfobject to refer to both parental behavior and a
"milieu" provided by parental behavior. This dual usage is confusing,
and neither is an intrapsychic usage of the term.
Kohut's treatment approach also contradicts his contention that self
object is an intrapsychic term. Recall that in Kohut's view the selfobject
transference is inevitably disrupted by the empathic failures of the
analyst, which are disappointing to the patient. When these distur
bances occur, Kohut believed, analysts should look for the ways in
which they were unempathic to the patient. However, if the selfobject
is intrapsychic, the analyst should seek the reasons for the disruption
in the analysand's subjective experience. The clear implication of the
recommendation that analysts look for their empathic failure is that
the analyst as a selfobject is a real person who performs activities that
have an impact on the patient.
Heinz Kohut 299
According to Kohut, the result of a successful analysis of a narcis

sistic personality disorder is the ability "to evoke the empathic reso
nance of mature selfobjects and to be sustained by them" (Kohut,
1984, p. 66). If a mature selfobject can be evoked to empathic reso
nance and perform or fail to perform functions, it cannot be intra
psychic. To Kohut's credit, he was aware of the criticism that his
usage of selfobject was confusing. His defense was that the reader
had to understand that when he used self-selfobject relationship he
meant the representations of "human surrounds" whether he
explicitly said so or not. This defense is not satisfactory because in
his statements regarding the origin of pathology, the selfobject com
ponent of treatment, and criteria for successful analysis, he used the
term selfobject to refer to behavior rather than "representatives of
the human surrounds."
Additionally, it is not at all clear that people without selfobjects
necessarily lead impoverished lives. Indeed, people differ a great deal
in their ability to withstand disapproval, unpopularity, and criticism.
Some people are much more needful of others' approval and are will
ing to compromise themselves to get it. To say that people need self
objects like they need oxygen is to oversimplify and mask crucial
psychological differences among people. In this sense, one may say
that the ability to live without selfobjects is a hallmark of self-esteem
and self-cohesion. There is considerable evidence that people need
both ties to others and self-enhancement, what Bakan (1966) calls the
"duality of human existence." Bakan marshals evidence to demon
strate the need for both "agency" and "communion," and social psy
chological evidence tends to support this view (Carson, 1969). While
Kohut understandably objected to the overemphasis on agency in the
psychoanalytic value system, he overreacted to it and neglected the
value of self-enhancement that is independent of echo, approval and
admiration and even in opposition to them.
The failure to recognize individual differences among people in the
extent to which they need others is a problem not only for the selfob
ject concept; it is a major weakness of Kohut's self psychology in gen
eral. By fitting all psychopathology into the category of self defects,
Kohut homogenized emotional disorder both with analytically
untreatable and with analytically treatable disorders. In the former
category, which is less germane to his major ideas, he blended border
line disorders with psychotic conditions in arguing that the former are
"really schizophrenic" but able to cover their psychosis with defense.
This view is a serious, almost glib, distortion of borderline psy
chopathology, which, as those who work extensively with borderline
300 Chapter 6
patients know, is not a psychosis but a stable, severe personality dis

order that is characterized by brief psychotic episodes under certain
conditions. We saw ample evidence of this concept of the borderline
personality in Kernberg's work (discussed in chapter 5), and this
conception of the disorder is substantiated by most workers in the
field of borderline pathology (for exam ple, M asterson, 1976;
Giovacchini, 1979; Adler, 1985). Why Kohut, who by his own admis
sion had minimal clinical experience with this group, would main
tain that borderline patients are "really schizophrenic" is difficult to
discern. More important, Kohut is guilty of homogenizing all ana-
lyzable pathology into one type of problem: defects in the self. The
treatment for all disorders is the same; namely, mobilization of the
archaic self and interpretation of narcissistic injury, which promote
transmuting internalization. This problem is similar to the difficulty
we encountered in the work of Fairbairn and Guntrip, who homoge
nized all pathology in terms of the schizoid position. While Kohut
has undoubtedly made a lasting contribution to the field with his
emphasis on and understanding of self pathology, his single-minded
approach risks losing the uniqueness of each case by assuming the
existence of a self defect that must be filled in by the analyst.
Alcoholic, depressive, obsessive-compulsive, hypochondriacal, and
phobic patients, to name but a few, are all in danger of being under
stood as suffering from the same deficit. This approach risks miss
ing those crucial aspects of each case that do not fit the model, the
very consequence Kohut criticized in what he saw as the intent of
classical analysis to make the patient fit the theory. Kohut's views
themselves are in danger of repeating the same mistake.
This single-minded focus of treatment leads to the third major diffi
culty with Kohut's views. Kohut was quite sweeping in his enuncia
tion of his principles of treatment. He believed the analyst should
allow the archaic self to become mobilized and should interpret only
failures and injuries. All disruptions are interpreted as the patient's
experience of the analyst's failures. This technical approach is far too
sweeping and potentially detrimental for patients who become "dis
rupted" and rageful for reasons other than the analyst's empathic fail
ures. For example, Kohut's treatment model would be difficult to
apply to patients who need to do battle and seek only the opportunity
to do so; it is equally difficult to apply to patients who project their
negative thoughts and feelings onto others, including the analyst.
Undoubtedly, one could fashion a self-psychological explanation for
such cases based on the empathic failures of the past and the patient's
need to safeguard self-cohesion. However, to interpret these situations
Heinz Kohut 301
as the patient's "phase-appropriate" response to the analyst's failure

does scant justice to the patient's pathology and risks submerging it
by putting the focus on the analyst's supposed mistake. Many patients
foster disruption and need to view others negatively in order to avoid
more deeply felt anxieties. It is difficult to see how the patient's
motive for the disruption and rage would be uncovered by an
approach that focuses on the analyst's empathic failure to meet the
patient's phase-appropriate needs. This is not to say that Kohut's
approach is not valid in many clinical situations; it is only to say that
his model is in danger of becoming a theoretical straitjacket that blurs
critical clinical differences. In this sense, self psychology may be erect
ing another version of what Goldberg (1990) calls the psychoanalytic
"prisonhouse."
Having said this, it should be noted that Kohut was responding
to the traditional analytic posture of fostering regression and then
criticizing the patient, even if subtly, by attacking resistances and
adopting a moralizing, educational posture. Kohut was unalterably
opposed to viewing patients' narcissistic needs as pathological entitle
ment and felt that analysts should adopt an empathic posture of
accepting and understanding regression as an appropriate response to
a given phase of pathology. While this approach has a great deal of
validity, it cannot be used to justify a single-minded focus on the ana
lyst's errors; part of a transference regression may well be the patient's
need to cause disruption of the analysis and be rageful toward the
analyst. It may be that Kohut's sweeping generalizations are an over
reaction to the classical model that blurs the distinction between fail
ures in analytic empathy and the patient's pathological needs for
disruption.
One can question whether Kohut's views risk losing the value of
transference analysis. In his view, the patient reacted in a realistic way
to the parents in the past and reacts in a similar fashion to the analyst
now; the current reaction is to be accepted as phase appropriate.
When the analyst misunderstands, the patient's rage is accepted as
understandable and, if interpreted at all, is seen only as an appropri
ate response to the analyst's "failure." This view seems to miss the
most powerful therapeutic aspect of transference: the patient's equa
tion of today's perceptions with past trauma. If the analyst is able to
point out to the patient that whenever the analyst acts in any way sim
ilar to a figure from the patient's past, the patient reacts as if to the
archaic figure, the analyst can help the patient understand the reaction
while accepting its appropriateness in the past. The patient's transfer
ence reactions frequently are responses to the analyst's behavior, but
302 Chapter 6
the therapeutic action lies in the patient's recognition that his or her
reaction is rooted in the past. This distinction is the most significant
and therapeutically powerful component of transference and is crucial
to helping patients understand the difficulties they have with many
people in their current life. If analysts cannot differentiate past from
present because they feels that their behavior is equivalent to that of
the childhood figure, they are colluding with the patient's pathologi
cal perceptions and are not using the leverage of the transference to
help the patient differentiate present echoes from past trauma and
thereby master his or her current emotional life.
It has fallen to Kohut's followers to attempt to complete his mission
by correcting problems in the model and continuing to develop it as
the primary alternative to classical psychoanalysis. To complete our
discussion of Kohut's work, we now turn to these followers to trace
recent developments in the theory of self psychology and thereby
grasp the state of the model at the present time.
THE SELF PSYCHOLOGISTS
Kohut's followers may be grouped into three broad categories: those

who adopt his self psychology model and make only minor emenda
tions to it; those who extend his model; and those who use Kohut's
work as a beginning point for a new concept of psychoanalysis even
more radical than self psychology.
Most of Kohut's followers fit into the first group. Wolf collaborated
with Kohut on the delineation of disorders of the self in order to refine
distinctions between different types of self pathology (Kohut and Wolf,
1978). They outlined four types of self pathology: (1) the understimu
lated self, which is bored and apathetic owing to the lack of selfobject
responses in development and which seeks excitement by pathological
means, such as sexual promiscuity, addictions, or perversions; (2) the
fragmenting self, which has lost its continuity in time and space owing
to the failure of selfobject responses and which resorts to hypochondri
asis when selfobjects fail; (3) the overstimulated self, which avoids
attention and receives no joy from success owing to its archaic fantasies
of greatness, caused by extreme lack of empathy or phase-inappropri
ate selfobject responses and which cripples productivity; (4) the over
burdened self, which suffers from diffuse anxiety and constantly
maintains a hostile view of the world and reacts to selfobject failures
with paranoia, pathology attributable to the inability of the childhood
selfobjects to allow merger and provide self-soothing functions.
Kohut and Wolf also delineated three personality types that are not
Heinz Kohut 303
necessarily pathological: the mirror-hungry personality, the ideal-

hungry personality, and the alter-ego personality. These personality
styles are pathological only if they are extreme, reflecting a deep defect
in the self. There are two such pathological personality styles: (1) the
merger-hungry personality, which has such an intense need for merger
that it lacks boundaries and is so intolerant of others' independence
that it attempts to dominate others, and (2) the contact-shunning person
ality, which has the same intense needs but reacts to them by avoiding
others. With these typologies, Kohut and Wolf hoped to differentiate the
various types of pathology that Kohut (1971) grouped under the rubric
"narcissistic personality and behavioral disorders" in his earlier work .
Wolf (1988), who prefers the expression "disorder of the self" to
"narcissistic disorder," sharpened Kohut's transference typology by
eliminating "mirror transference in the broad sense." Instead, he
divided selfobject transferences into the following types: merger, mir
ror (formerly "mirror in the narrow sense"), alter-ego, ideal, and
adversarial. He defines the latter as the expression of the patient's
need for supportive, oppositional selfobjects that is, like other selfob
ject needs, lifelong. Wolf also attempted to define the selfobject needs
of the life cycle, pointing out that there are, in addition to the archaic
and oedipal selfobject needs discussed by Kohut, selfobject needs in
the more advanced stages of the life cycle: in latency, selfobjects are
needed as models for imitation whereas in prepuberty the need for
selfobjects moves away from parents and toward teachers, friends,
and symbols. Adolescents and young adults find selfobjects in peers,
idols, and the subculture; and in marital relationships the spouse
becomes the primary selfobject.
Apparently aware that Kohut's view of selfobjects as psychological
oxygen can be criticized for failing to take into account people who
appear to function well without approval or admiration—or even
without others—Wolf points out that Kohut (1984) had extended the
selfobject concept to activities, such as listening to music and reading.
Wolf believes that selfobject needs are concrete only in infancy. As the
self develops, selfobject needs become increasingly abstract: symbols
or ideas can serve the purpose. Activities and ideas are symbolic
replacements for the selfobjects of infancy. This view may explain
why some people appear to function without selfobjects, since differ
ent experiences might work for them, but it creates new problems.
First, Kohut defined selfobject in general as the "human surrounds"
and more specifically as the echo, admiration, and approval of the
human environment. If music and reading are categorized as selfob
jects, one would have to say that they somehow serve to admire and
304
approve. Since they clearly cannot do this, the selfobject concept has
to be expanded to mean anything that makes people feel good. Thus,
ultimately it can mean and explain nothing. In this broadened sense
the selfobject concept can have no explanatory power; it can only label
after the fact any positive experience as "selfobject." This does not
help to understand individuals, nor does it advance the field. In this
broadened sense the concept of selfobject is no different from the
concept of reinforcement in behavioral theory; it simply labels what
people do by means of a pseudoexplanatory theoretical construct.
Self psychologists have tended to underscore the value of empathic
resonance in the therapeutic process. Basch (1985) considers "affective
attunement" the paradigmatic self-selfobject relationship, applicable
to the mother-infant dyad as well as the therapist-patient relation
ship. According to Ornstein and Ornstein (1985), the understanding
phase of treatment constitutes acceptance of the patient's childhood
wishes as legitimate, an experience that leads to belated maturation
and growth. P. Tolpin (1988) considers the therapeutic action of psy
choanalysis to be in the "optimal affective engagement" of the patient,
a phenomenon tantamount to the therapeutic provision of selfobject
functions. M. Tolpin (1983) elaborates Kohut's concept of the "correc
tive emotional experience": her view is that the mutative aspect of the
analytic process lies in the working through of a new "transference
edition" of the old self-selfobject unit. The interaction between patient
and analyst, once the selfobject transference is established, is equiva
lent to a "corrective developmental dialogue" in which the patient's
developmental needs are responded to in a new way; it is the patient's
internalization of this "cohesion-fostering" selfobject tie that is the
essence of the curative process. For Tolpin, who emphasizes this new
experience as the essence of the change process, this is the self-
psychological "corrective emotional experience."
One can see that those of Kohut's followers who apply his concepts
closely tend to de-emphasize the "optimal frustration" aspect of
insight in favor of a new "exchange" or experience between analyst
and patient. This tendency is even more prominent in the second
group of self psychologists, who tend to accept most of Kohut's prin
ciples but further shift the treatment focus from insight to the
patient-analyst relationship. This view is best represented by Bacal
and Newman (1990), who are in the minority among self psycholo
gists both in their acknowledgment of the debt of self psychology
to previous object relations theories and in their view of self psychol
ogy as one type of object relations theory. These theorists agree with
Wolf and other self psychologists that the crucial contribution of self
Heinz Kohut 305
psychology is the selfobject concept. Unlike other self psychologists,

however, they believe that self psychology is a "multi-body" theory
because the self-selfobject relationship is an interpersonal relation
ship. Since the self is formed from the psychological meaning that
grows out of self-selfobject interactions, the mutative aspects of the
analytic process are considered to lie in the provision of selfobject
functions. Bacal and Newman list five such functions: affective attune-
ment, validation of subjective experience, tension regulation and
soothing, organization and restoration of the self, and recognition of
uniqueness. The task of the analyst is to provide these functions, and
to the extent that he or she does, the analysis will be successful.
Bacal (1985) replaces Kohut's concept of optimal frustration with
the concept of optimal responsiveness. According to Bacal, there is
no inherent value in frustration and Kohut's use of the concept is a
vestige of the drive model that is best removed from his theory.
According to Bacal, if the child is responded to optimally, its self will
grow; if not, growth will be delayed. Bacal and Newman point out
that since Kohut rejected the view that the "untamed" infant must
experience frustration in order to turn drives into structure, optimal
frustration can have no role in his therapeutic model.
The upshot of the shift from optimal frustration to optimal respon
siveness is to exchange the role of insight in the treatment process for
the provision of selfobject functions. Bacal and Newman believe the
patient needs to be responded to, not investigated. The value of expla
nation, in their view, is to verbalize affects—much as the mother
names the child's experience to help the child organize it. However, if
the analyst's only way of communicating understanding is to put feel
ings into words, the patient's experience may be invalidated, as
patients cannot always use words. The analyst's way of talking, listen
ing, and responding may provide the needed selfobject experience
more effectively. It is this view of the analyst's task—that is, to pro
vide selfobject functions rather than insight—that makes Bacal and
Newman's formulations an extension of the self-psychological model
rather than a refinement of it. For Kohut, the analytic process consists
of the provision of insight, although its value lies not in the accretion
of knowledge, but in the creation of new self-structure. For Bacal and
Newman, the analytic process involves the provision of self object
functions, which may or may not include insight.
There are two offshoots of self psychology that represent an even
more radical departure from the drive-ego model. The first is the
intersubjective model advanced by Stolorow and his followers
(Stolorow and Lachmann, 1980; Stolorow, Brandchaft, and Atwood,
306 Chapter 6
1988), who point out that Kohut's definition of psychoanalytic data

as that which is gathered by empathy and introspection makes psy
choanalysis a "pure psychology" (Stolorow, 1985). Stolorow and his
colleagues agree with Kohut that drive is not a psychoanalytic con
cept because it is not accessible to the psychoanalytic method. They
argue that once psychoanalytic theory is freed from the drive con
cept, one is left with "intersubjective contexts." According to Kohut,
the traditional analyst cannot be empathic with the narcissistic
patient because the material does not fit the analyst's preconcep
tions. Stolorow and his colleagues call this an example of the ana
lyst's subjectivity failing to connect with the patient's subjectivity.
Kohut believed he made progress with Miss F when he gave up his
attempt to fit her material into the model of the Oedipus complex
and heard her expectations as longings to meet a developmental
need; Stolorow points out that such a clinical experience indicates
that the analytic process is a meeting of two "subjectivities." He
believes his intersubjective theory frees psychoanalysis from the last
vestiges of the drive-ego model. The analyst's task is to understand
the patient by attempting to grasp the material from the patient's
point of view, the patient's subjective reality. Analysts have their
own viewpoint from which they will interpret the patient's reality,
but Stolorow's point is that the analyst's reality is no more objective
than the patient's view and that it therefore has no claim to priority.
He is supported in this view by Schwaber (1979), who sees "analytic
listening" as the grasp of the patient's "subjective truth."
From the vantage point of their more radical phenomenological
view, Stolorow and his colleagues agree with Kohut's emphasis on the
centrality of self experience and its relationship with selfobjects, but
they find fault with Kohut's concept of the self and its development
(Stolorow et al., 1988). First, they point out that Kohut uses self to
mean both the sense of personal agency and a structure that organizes
experience. The first meaning confuses self with person. Second, they
view the bipolar self as a simplification and reification. In Stolorow's
view, ambitions and ideals are two affective meanings in a multidi
mensional self with a variety of affective meanings; moreover, the
"tension arc" that allegedly spans them is not accessible to empathy
and introspection. Third, Stolorow and his colleagues object to the
notion that the self can break down, and have "disintegration prod
ucts." They believe that such concepts are reifications, remnants of
the old mechanistic metapsychology, and serve only to obscure the
meaning and purpose of affects.
One can see from this critique that, in the Stolorow group's view,
Heinz Kohut 307
the self is organized around affects and the understanding of affects

is crucial to the analytic process. To achieve its organization of
affective meanings, the self requires the responsiveness of care
givers. Stolorow and his colleagues (1988) point out that the concept
of optimal frustration is based on the drive-frustration model of
development and that "transmuting internalization" is a mechanis
tic concept inaccessible to empathy and introspection. They agree
with Bacal that the self requires responsiveness, not frustration.
Stolorow and his coworkers replace the concepts of optimal frustra
tion and transmuting internalization with the hypothesis that the
self develops through integration of its affective meanings. Self
objects, in their view, are functions that help the self to integrate
affects into the organization of self-experience.
Stolorow and his colleagues (1988) describe four key self object
functions performed by the parents: affect differentiation, the synthe
sis of affectively discrepant experience, toleratation of affects and their
use as signals, and the "desomatization" and logical articulation of
affects. If any of these selfobject functions are not performed in early
development, the child's central affective states are not effectively
responded to; as a result, they must be disavowed rather than inte
grated, and development is derailed. Such a self is vulnerable owing
to its lack of integration of affects.
Although Stolorow tends to eschew diagnostic labels, clinical
depression can be invoked to illustrate his viewpoint: if the child's
depressive affects are not responded to in childhood, these affects are
not integrated into the self structure; when grief reactions or depres
sive affects of any type are subsequently experienced, the self is
threatened with dissolution. According to the Stolorow group, this
threat is the basis of all psychogenic affective disorders. The crucial
variable in the onset of affective disorder is the lack of calm caregiver
responsiveness evoked by the depression.
If the parent is unresponsive but the child adapts to the parent to
maintain the selfobject tie, guilt conflicts result (Stolorow, 1985).
According to Stolorow, conflicts involving guilt are a product of
what the child must do to maintain the tie to parents who are
unable to adapt to the child. In this view, the superego is the child's
perception of what is required to maintain the selfobject tie. If the
child's autonomy is a threat to the parent, the child will feel that its
own affects are threatening and will believe itself to be cruel and
dangerous. This is Stolorow 's concept of what Kernberg calls
"superego forerunners" (see chapter 5). Structural conflict is always
a product of selfobject failure.
308 Chapter 6
In the view of Stolorow and his colleagues, more severe pathology

has the same root, but the injury is so severe that affective strivings
are abandoned in favor of self protection to prevent repeated trauma.
Either the self is protected by rebellion or isolation or it submits and
becomes chronically dependent; or it is tormented by ambivalence
between these two states. Brandchaft and Stolorow (1984) view terms
such as borderline as pejorative labels that reflect the clinician's failure
to understand the patient's developmental needs. In keeping with
their intersubjective approach, they contend that borderline pathology
does not reside in the patient but is always codetermined by the inter
subjective context. In such severely disturbed patients, it is difficult to
comprehend the archaic intersubjective contexts in which the pathol
ogy arises, resulting in recourse to terms like borderline. The therapeu
tic process with such patients consists of the same process as Kohut
described in the treatment of narcissistic disorders, but the "archaic
intersubjective meaning context" is more difficult to find.
In accordance with this theoretical posture, the goal of psycho
analysis for Stolorow is the unfolding, illumination, and transforma
tion of the patient's subjective world. Consequently, the task of the
analyst is to facilitate the unfolding of developmental strivings. The
stance of the analyst is determined by whatever helps achieve this
goal; if abstinence does not facilitate this goal, it should not be part of
the analytic stance. Stolorow believes that the optimal approach to
facilitate the transformation of the patient's subjective world is sus
tained empathic inquiry. Such an attitude permits the structure of the
past to gain expression in the analytic relationship. From this view
point, the transference is the organizing activity by which the patient
subsumes the analyst into his or her view of the world. Stolorow
objects to the concept of a temporal regression. Instead, he points to
a universal striving to organize experience and understands psy
chopathology as the operation of the patient's past "structures of
subjectivity" in the present.
Stolorow and his colleagues adopt Kohut's view that failures in self
object experience are pathogenic, but they believe that conflict has a
larger role in emotional disorder than Kohut ascribed to it. While
Stolorow agrees with Kohut that selfobject failure results in develop
mental arrest, he points out that when the self is able to resume devel
opment by the establishment of a selfobject tie, such as occurs in
analysis, the self will then be in conflict. Treatment may involve pro
longed periods of empathy and understanding to establish the selfob
ject tie, but once it is so established, conflict can be addressed; when
this bond is disrupted, deficits become prominent and conflict recedes.
Heinz Kohut 309
Deficit and conflict are figure and ground, according to Stolorow, and
each becomes the focus at different times in the treatment.
The therapeutic action of analysis in this view is the mutative effect
of the selfobject tie. Initially, the patient is defensive out of fear of a
faulty response from the selfobject. Once this resistance is analyzed,
the emergence of the need for the original selfobject allows for the
articulation of the patient's affective organization in the analytic set
ting. As the analyst becomes a part of this organization, the transfer
ence bond is established, and to the extent that it is protected from
disruption, growth can occur. When disruptions take place, they are
interpreted, thereby permitting the reestablishment of the tie and the
resumption of arrested development. According to Stolorow, it is this
transference tie that gives interpretations their mutative power; inter
pretations are part of this bond. The selfobject tie and transference
cannot be separated. That is, in this view, all mutative moments in
analysis entail some degree of "transference cure." This process is as
applicable to the treatment of the so-called borderline as it is for the
patient with structural conflict. Insight, affective bond, and psycholog
ical integration are critical components of every psychoanalysis. The
resolution of this process occurs when the transference experience
becomes integrated into the patient's psychological organization.
When this happens, the patient's affective life is immeasurably
enriched.
The other radical offshoot of self psychology is the approach of
Goldberg (1990), who uses Kohut's concept of selfobject as the link
between psychoanalysis and the modern philosophical and scientific
view of man. In Goldberg's view, traditional psychoanalysis is still
wedded to the outmoded subject-object distinction. That a person is
not limited by his or her skin is, according to Goldberg, a conclusion
of many disciplines. For example, memory is not a static "storehouse"
but an active effort, a working upon the world; external objects acti
vate neurons and the connections between neurons so that the object
actually becomes part of the entire pattern of activating the brain.
There is no storehouse of internal representations upon which people
draw; what we call representations are ways of participating in the
world. Heidegger's philosophy of man as being-in-the-world high
lights the inherent, inextricable link between man and world.
In Goldberg's view, Stolorow's conception of separate "subjectivi
ties" coming together is an advance but does not go far enough
because it fails to appreciate that one's "subjectivity" is never sepa
rate, that there is no need to postulate a "coming together." Gold
berg adopts Heidegger's view that we are inherently bound in a
Chapter 6
man-world relationship, and our particular relationship to the world

and others is who we are. According to Goldberg, Kohut's concept of
selfobject is the psychoanalytic dimension of this modern view of
man. The implication of Kohut's view of the self-selfobject relation
ship is that the self is composed of the experience of others, selfobjects.
Goldberg's view is that among psychoanalytic theories only Kohut's
view of the self-selfobject relationship appreciates the inherent rela
tionship between man and world. Freed from the outmoded self
object distinction, analysis can, Goldberg contends, become free to
study the development of the self as it changes its composition by
means of its self-selfobject relationships. In this view, transference is a
form of Heidegger's being-in-the-world. The psychoanalytic process
then becomes an effort to transcend the boundaries of the self, to find
new ways of being. According to Goldberg, the essence of psycho
analysis is a mutual effort between patient and analyst to remake the
patient's self via a verbal exchange. In Goldberg's conceptualization of
the process, self-psychological concepts become a bridge between
psychoanalysis and the phenomenological-existential view of man.
By way of summary, one can see that although the various off
shoots of self psychology differ and even conflict in some areas, they
have all tended to move self psychology further away from its roots in
the drive-ego model. In this sense, self psychology is evolving toward
a fundamental shift in the psychoanalytic conception of the nature of
the human enterprise and its requirements for change. Even the more
conservative tendencies of the self psychology movement have
eschewed the drive model remnants in Kohut's theory in favor of
emphasis on "affective engagement." Self psychologists whose theo
ries are more sharply divergent from the classical model have shifted
toward a relationship paradigm of psychoanalysis in opposition to an
interpretive model of treatment. Further, the most radical branches of
self psychology have used Kohut's conceptual reformulation of psy
choanalytic thought as a springboard to a pure psychology of people-
in-interaction with the world. This paradigmatic shift is the link
between object relations theories such as self psychology and interper
sonal theories of psychoanalysis, and it is to this latter school of
thought that we now turn.
CHAPTER 7
The Interpersonalists
The in t e r p e r s o n a l a p p r o a c h t o p s y c h o a n a l y s is w a s d e v e l o p e d
between the 1930s and 1950s by Harry Stack Sullivan but was largely
ignored by mainstream psychoanalysis until the growing influence
of object relations theories led to recent interest in integrating it with
more traditional models. A significant component of this shift is
Greenberg and Mitchell's (1983) interpretation of object relations
theories in terms of a "relational model" conceptually linked to
Sullivan's interpersonal theory. Let us delineate the major concepts
of interpersonal theory, ascertain its similarities to and differences
from object relations theories, and assess its contribution to psycho
analytic thought. Since the interpersonal theory of emotional disor
der began with Sullivan and all current versions of this model are
influenced, if not rooted in, his theory, our discussion begins with
his contribution.
THE WORK OF HARRY STACK SULLIVAN
Sullivan found Freud's metapsychology of impulses and psychic

mechanisms to be too far removed from patients' experiences. The dif
ficulties people have in living have to do with their interpersonal rela
tions, he believed, not intrapsychic mechanisms. Sullivan's (1953)
view is that human beings are inherently imbedded in such relation
ships to the point that one cannot understand them outside their rela
tionships with others. The way one relates to others defines who one
is. In Sullivan's view, this fact can be seen from the beginning of life.
The baby cannot be conceived of without the mothering figure. And,
however independent children may become, they define themselves
by the way they relate to others. Freud's intrapsychic concepts—
impulses, defenses, ego, superego, id, even the unconscious—only
obfuscate understanding by artificially cutting off individuals from
their relationship to the human environment.
In accordance with his concept of the nature of personality,
Sullivan (1953) conceived of development as a series of stages of inter
personal relationships. The infant is born with "needs for satisfac
tion," which include both biological and emotional needs. A need
311
312
produces tension, but the infant is completely incapable of satisfying

the need by itself. The infant's tension elicits tension from the mother
ing figure, which leads her to satisfy the need. The mother's ministra
tions are experienced by the infant as tender behavior, and "these
needs, the relaxation of which require the cooperation of another,
thereupon take on the character of a general need for tenderness"
(Sullivan, 1953, p. 40). That is, the generic needs of the child for the
mother are subsumed under the need for tenderness, which becomes
a primary emotional need. Even though many of the infant's needs
involve biological tension, they are considered interpersonal because
they can only be satisfied by tender behavior from another. Sullivan
called this first relationship "maternal empathy" and believed that it
characterized the first stage of development.
Life remains interpersonal as the infant develops because needs
continue to require others for their satisfaction. In the first year the
infant's emotional need is for bodily contact. Between ages one and
four, the need is for others to be an "audience" for the child's behav
ior. For children ages four through eight the primary need is for learn
ing both to compete and compromise with others. The primary need
of the preadolescent, ages eight to puberty, is for a "chum" of the
same sex with whom to be intimate, and in adolescence the need is for
intimacy with the opposite sex. Each of these stages is defined by the
type of relationship sought.
One can see from this view of personality development that
Sullivan attributed much more importance to emotional needs than
to biological needs. He acknowledged the pressure of biological
tensions, especially in early development, but he pointed out that
since these tensions immediately give rise to the need for another
person, the need for tenderness predominates. It is the history of
relations with the mother and then other persons that forms the
personality. However, needs for satisfaction are not the only type
of need. If the infant's needs elicit anxiety in the caretaker, the
infant "catches" this anxiety and becomes fearful. Anxiety, for
Sullivan, is different from tension because it has no identifiable
object; it is difficult to describe in words but approximates an
uncanny sense of dread. Sullivan's primary theorem regarding anx
iety is that it is contagious, that is, it is exchanged through the
empathic linkage between caretaker and child. When the child's
needs produce tenderness in the mothering figure, the needs are
met and the tendency is toward integration of the personality;
when the needs produce anxiety in the caretaker, the child becomes
anxious and the experience is "disintegrating."
The Interpersonalists 313
According to Sullivan, the child's first distinction is between anxious

and nonanxious states, and these global, diffuse conditions are experi
enced as "good mother" and "bad mother." These "prehensions" are
not perceptions of the real mother; rather, all experiences of tenderness
and need satisfaction, from however many people, are "good mother"
while all experiences of anxiety and disintegration are "bad mother."
Further, self- and object-images are not differentiated at this point. Only
later, as the child matures, do these images and the particular character
istics of self and other become differentiated. As this differentiation
occurs, the child develops a view of itself based upon its perceptions of
the way it is being viewed and responded to by its caretakers. Sullivan
terms these perceptions "reflected appraisals." The self is defined as
what one "takes oneself to be," and this sense emerges from the inter
personal patterns of development. All experience consistent with oth
ers' appraisals becomes part of the self whereas experience diverging
from these appraisals is excluded from this organization.
According to Sullivan, anxiety is the most unbearable feeling to
which the human being is subject. Once anxiety is felt, the personality
develops a strong need to avoid it. The need to avoid anxiety becomes
the "need for security," the second basic human motivation. Because
of its fear of anxiety, the child learns to anticipate which of its behav
iors will produce anxiety in the caretaker and to pattern its interac
tions accordingly. Sullivan called the maneuvers used by the child,
and later by the adult, to avoid anxiety "security operations." The ini
tial security operations are means of restricting awareness. Behavior
that gains caretaker approval, increases tenderness, and reduces anxi
ety in caretaker and child becomes organized into the "good me."
Behavior that has the opposite effect, that results in the mothering fig
ure's disapproval, reduces tenderness, and increases anxiety in both
the child and caretaker, becomes organized into the "bad me."
Sullivan added a third category of experience that produces such
intense anxiety in caretaker and child that it cannot be allowed into
awareness in any form; he called this experience the "not me or disso
ciative system." Experience of this type operates on the edge of aware
ness but is occasionally experienced by all people, the most common
instance being nightmares. When "not me" experience breaks into
awareness, it tends to be felt as dread or uncanny emotion. The per
sonality attempts to keep all "bad me" and "not me" experiences out
of awareness. Eventually, the mother is no longer necessary, as the
self-system is sufficiently developed to reduce anxiety by restricting
awareness only to what fits. As the child matures, it learns other forms
of security operations, that is, other ways of reducing anxiety, most of
314
which involve a sense of superiority. The primary maneuvers dis

cussed by Sullivan are the illusions of power, stature, and a sense
of specialness. The self-system for each individual is the set of secur
ity operations that effectively reduced anxiety in childhood, and
preserves the shape of the self.
Anger, in Sullivan's view, is a response to threat. The child's first
response to punishment is rage, but as it develops and learns foresight
it finds that anger is more useful. If anger is suppressed, the child may
continue to be rageful and have temper tantrums into the school
years. If the child is punished with no discernible cause, it will learn
that anger exacerbates the situation and will become chronically re
sentful. Resentment, in Sullivan's view, is anger that is not expressed
openly for fear of reprisal. Most children learn anger and many are
resentful to some degree, but the most debilitating form of anger
is what Sullivan called the "malevolent transformation." This patho
logical shift in the psychological organization comes about if the
child's need for tenderness is not only denied but met with a hurtful
response. If children are made to feel anxious or humiliated for the
expression of their need for tenderness, they will turn this need into a
malevolent attitude and not only refuse to show such a need but
refuse to allow anyone to act tenderly toward them. In these cases
hostile behavior becomes a significant component of the personality
and greatly interferes with interpersonal relationships. As can be seen,
Sullivan did not view anger or aggressiveness as inborn; he saw them
as reactions to situations of interpersonal anxiety. Anger is always a
response to threat, and since all people experience threats, they all
experience anger. It is not an emotion to be tamed, and it becomes
part of a pattern of emotional disorder only when it replaces the need
for tenderness.
Unlike most psychoanalytic theorists, both traditional and object-
relations-oriented, he did not accept Freud's theory of infantile sexual
ity. Adolescence begins what he called the "lust dynamism." Lust is
the need for satisfaction characteristic of this phase of development;
like all needs for satisfaction, it requires another for fulfillment. How
well it becomes integrated with the need for intimacy is a function of
the interpersonal experiences of this phase. The degree to which these
experiences are marked by anxiety is the degree to which the lust
dynamism becomes a disintegrating rather than an integrating ten
dency in personality development. Like anger, sex is not a drive to be
tamed but an interpersonal need that becomes growth enhancing to
the degree that it elicits tenderness as opposed to anxiety.
Human life, in Sullivan's view, revolves around the dialectic
between the dual needs for satisfaction and security. When a need is
associated with minimal or no anxiety, it will be satisfied; when anxi
ety interferes, the need for security interferes with satisfaction. The
degree to which the need for security dominates the need for satisfac
tion is the degree of psychopathology in the personality. Since the
anxiety that motivates security operations always has an interpersonal
source, all psychopathology is ultimately traceable to anxiety-ridden
interpersonal relationships. A pathological personality is a self that
operates only, or primarily, out of the need for security. In such a per
sonality the needs for power, status, and prestige predominate.
Sullivan did not use traditional diagnostic categories, in keeping with
his belief that psychopathology is a human process, not a "disease"
that can be diagnosed. In his schema diagnosis is supplanted by an
assessment of the interpersonal sources of the patient's security opera
tions. All the traditional nosological categories, such as obsessive-
compulsive and hysterical disorders, are viewed as markers for dif
ferent security operations that were developed originally in child
hood in response to anxiety and are maintained out of fear of its
reappearance.
Negative or unhappy interpersonal experiences of childhood lead
to security operations, but if these operations cannot protect against
the experience of the "bad me," low self-esteem results. For Sullivan,
this low self-esteem is the source of many traditional psychiatric syn
dromes. Since the "bad me" produces anxiety, new strategies of anxi
ety reduction must be developed. Low self-esteem, in Sullivan's view,
inhibits what he called "conjunctive motivation," that is, integrating
situations capable of bringing need satisfaction, such as falling in love.
Consequently, other solutions are sought, resulting, for example, in
"exploitive attitudes," which include behavior that would normally be
classified as passive-dependent and masochistic. Sullivan equated the
general characterological recourse to masochism to an indirectly
exploitive use of others, as opposed to the more open efforts of
exploitation used by the passive-dependent character. In lieu of
exploiting others, individuals may tell themselves how abused they
are or may envy others, and Sullivan called these strategies "substitu
tive processes." He also considered hypochondriasis on this continuum,
as he believed it to be the personification of oneself as "customarily
handicapped." People with low self-esteem may also involve them
selves in situations in which they are taken advantage of in such a
way that they are led to expose a weakness. For example, they may
engage in relationships in which others entrap them into acknowledg
ing something they would normally keep private; the result is chronic
316 Chapter 7
humiliation. When the exposed weakness is part of the person's "dis

sociative system," the experience is "attended by some measure of the
uncanny emotions—awe, dread, loathing, and horror" (Sullivan, 1953,
p. 359). In Sullivan's view, this experience of the uncanny is the closest
one can get to the "not me," the dissociative part of the personality,
without becoming schizophrenic. If the individual is unable to employ
security operations against the previously dissociated component of
the personality, it becomes the personified "not me." Sullivan calls
this the "paranoid transformation," which in turn can degenerate into
a schizophrenic process.
Sullivan explained all of what would normally be considered
psychopathology in terms of various reactions to interpersonal
anxiety that he believed corresponded to real experiences in child
hood. This explanation encompassed even the most severe form of
pathology, schizophrenia (Sullivan, 1962). Indeed, Sullivan applied
much of his early theoretical and clinical effort toward advancing
the view that schizophrenia is an understandable reaction to inter
personal anxiety, like any other form of pathology. He believed
that schizophrenics had suffered such devastating anxiety in early
life that their security operations could not be employed effectively
against the dissociative component of the personality, and he
labeled the resulting terror "schizophrenia."
Thus, Sullivan attempted to account for all difficulties in living
without recourse to traditional psychoanalytic concepts, such as the
unconscious, the ego, the id, defenses, and even internal objects. He
substituted a dual system of human needs for Freud's dual drives and
attempted to explain psychopathology on the basis of the relative
balance between the need for security and the need for satisfaction.
Consistent with his interpersonal theory of development and
pathology, Sullivan (1956) conceptualized the goals of treatment as the
learning of more effective, flexible ways of interacting with other peo
ple. His focus was on what patients do in their interpersonal world in
the present and on what can be done about it, rather than on uncover
ing the past. Thus, Sullivan's (1940) goal with all patients was to make
them aware of their patterns of interpersonal relationships. Perhaps
his most commonly quoted statement is that "one achieves mental
health to the extent that one becomes aware of one's interpersonal
relations" (Sullivan, 1940, p. 102).
Sullivan's use of the transference is puzzling. He recognized that
the patient-therapist relationship is from a theoretical viewpoint, an
interpersonal relationship, as he defined the role of the therapist as
"participant observer." He believed that the therapist is inevitably a
participant in the interaction and that no useful data could come from
trying to adopt a "detached position." Sullivan (1954) adopted the
theoretical position that the therapist must begin with the patient's
external relationships and later move the treatment into the patient-
therapist relationship. Despite this view, however, there is considerable
evidence that he focused very little on transference and countertrans
ference. While he recognized them in the form of resistances and
asserted their importance, he in fact interpreted the transference very
little (Havens, 1976).
Sullivan's technique seems to have been based on two strategies.
The first is to help patients with their relationships outside of the
patient-therapist interaction by attempting to help them gain aware
ness of their interpersonal patterns. The second strategy, which
addresses the patient-therapist interaction, is the attempt to coun
teract, rather than interpret, the patient's perceptions. In his dis
cussion of psychiatric interviewing, Sullivan (1954) described a
variety of techniques, verbal and nonverbal, for demonstrating to
the patient that what the patient feels as shameful and anxiety pro
voking is not so construed by the therapist. In addition, Sullivan
made many statements derogating interpretation as a whole, not
just transference interpretations. Havens (1976) termed Sullivan's
technique "counter-projective," by which he meant that Sullivan
actively combatted the patient's projections. It appears that Sulli
van's attitude toward treatment was that patients need a new inter
personal relationship in which the other party is not as fearful of the
patients' emotions as the patients are themselves. Nonetheless,
Sullivan believed that the goal of treatment was the patient's aware
ness of the patterns of interpersonal relationships; one may infer
that he believed his techniques of counterprojection would help
the patient become more aware of these patterns, but this was not
made explicit.
The fact that Sullivan made little use of unconscious mental proc
esses and put little emphasis on interpretation in general or on trans
ference interpretation in particular largely accounts for the fact that
his thought received little attention outside Sullivanian circles for sev
eral decades, even among those favorably disposed to his interper
sonal theory. For example, Guntrip (1961a), who lauded Sullivan's use
of interpersonal concepts in lieu of drives, did not consider Sullivan
a psychoanalyst because his theory was not a depth psychology.
However, in recent years Sullivanian analysts, who adopt major tenets
of Sullivan's interpersonal approach to development and pathology,
have incorporated his theories into a system of psychoanalytic
318 Chapter 7
exploration and interpretation. These Sullivan-inspired models of the

psychoanalytic process cannot be summarily dismissed as "nonana-
lytic." It is to these more contemporary outgrowths of Sullivanian
analysis that we now turn.
THE WORK OF JAY GREENBERG AND STEPHEN MITCHELL
Greenberg and Mitchell (1983) view Sullivan's work as one of many ver
sions of the "relational/structural" model, the others being the various
object relations theories. All theories based on a view of man as inher
ently imbedded in human relationships fit this model in which is oppo
sition to the "drive/structural" model of classical psychoanalysis. That
is, Greenberg and Mitchell recognize two basic theoretical models of psy
choanalysis: one that sees man as an interpersonal being motivated
to relate to others as part of his very nature and one that considers the
fundamental human motivator to be inborn drives and their vicissitudes.
Greenberg and Mitchell recognized that some object relations theo
ries, such as Kernberg's, attempt to accommodate both models by
viewing object relations as the primary preoedipal issue and drive/
conflict as the appropriate model for the neuroses. They take the posi
tion that all such "strategies of accommodation" are doomed to fail: if
the child is viewed as inherently imbedded in a relational matrix, the
drives must be seen as fundamentally object relations and the drive/
structural component never quite works, even for the neuroses in
these hybrid theories. We saw that this was the case in the critique of
Kernberg in chapter 5, where it was pointed out that Kernberg's theo
retical model obviated the drive concept, although he himself had not
drawn this conclusion. Greenberg and Mitchell opt for what they term
a "strategy of radical alternative," replacing the drive model with the
"relational/structural" model, with Sullivan's theory as one of many
contributions to this model.
Mitchell (1988) has adopted the view that psychological reality is a
relational matrix encompassing both the intrapsychic and interper
sonal realms; he opposes this model to the "monadic view of the
mind," which assumes that the self can operate independently of oth
ers. He considers the principal contributors to the relational model to
be the interpersonal psychoanalysts, beginning with Sullivan and
Fairbairn, and considers object relations theorists who emphasize self
development, such as Kohut, to be "monadic" theorists because their
basic units, such as the nuclear self, are intrapsychic. Mitchell is also
critical of theorists like Kohut, Winnicott, and Guntrip, who viewed
pathology as developmental arrest, for failing to appreciate both the
inherent nature of human conflict and the relational nature of develop-

ment. Mitchell's alternative is a "relational conflict" model, which rec
ognizes the intrinsic nature of human conflict but sees its basic units as
relational configurations, not drives and defenses against them. For
Mitchell, mind does not need to become socialized: it exists only as a
social product. Therefore, the units of psychoanalytic study are rela
tional bonds and the matrices they form. One can see in this view the
influence of Sullivan's interpersonal theory of mental life. However,
unlike Sullivan, Mitchell is concerned with the meaning of experience
to the individual and sees psychoanalysis as the process of elucidating
meaning, especially as it manifests itself in the patient-therapist inter
action. Thus, Mitchell draws out the clinical implications of Sullivan's
interpersonal theory by proposing that meaning is embedded in a rela
tional matrix and that the relationship between patient and analyst is
the best place to discover the meaning of the patient's experience.
In accordance with his substitution of the relational model for the
drive model, Mitchell does not view sexuality as a drive. He appreci
ates its biological "drivenness" and acknowledges its prominence in
human life but points out that sexuality always gains expression in a
relational matrix. Thus, sexuality does not derive its power from
organ pleasure but from its meaning in a relational matrix. Although
sexuality is an especially powerful motive, it is a vehicle for the ex
pression of relational patterns and social roles. It does not create
social meaning but derives its power from the social meaning with
which it is relationally invested. Mitchell believes that sexuality is
particularly apt to become the focus of psychopathology because
its biological power and requirement of another makes it the most
powerful medium for connection with other people and renders
the individual potentially vulnerable to the needed other. Further,
the privacy of the body makes sexuality a natural vehicle for the
symbolization of object seeking.
If development has resulted in object seeking becoming a danger
ous endeavor, sexuality becomes not an enriching experience, but "a
search for symbolic reassurances and illusory guarantees" (Mitchell,
1988, p. 111). For example, the search for the elusive object may be
concretized by the genitals, resulting in compulsive promiscuity. If
demands from the primary object to be good and clean were exces
sive, sex may be associated with dirtiness; in an extreme outcome,
only prostitutes or other "dirty" or "degraded" objects are exciting
because they represent defiance of the primary object. Or, if submis
siveness is felt to be a requirement to maintain connection to the part
ner, sexuality with the primary object may become deadened and
320 Chapter 7
sexual dominance in fantasy or reality with other objects may become

the only means for independence. Mitchell gives other such examples
to make his point that sexual difficulties of any form are sexualized
expressions of relational conflicts.
Mitchell is critical of object relations theorists, such as Kohut and
Winnicott, who conceive of the patient as a passive receptor rather
than a creator of experience and who reify the patient's experience of
himself or herself as a baby with their belief that patients who have a
concept of themselves as a baby are arrested at an infantile life that
lurks under the surface of adulthood. Mitchell views such a self as a
strategy that permits interaction with others. Further, he regards relat
edness as inherently conflictual and believes that the self-as-baby is a
product of conflict rather than an unconflicted arrested self that needs
only to be found. In addition, Mitchell believes that the needs
expressed by the adult patient are not infantile needs but adult depen
dency needs fraught with anxiety. All human needs exist throughout
the life cycle, and the developmental arrest hypothesis, in his view,
conflates ongoing life cycle needs with their earliest expression.
In contradistinction to the developmental arrest view, Mitchell
believes that patients willfully cling to their pathological patterns as
the only types of relationships they know. According to Mitchell, all
children have a range of possibilities and this range is limited by the
parents not because of parental failure but because the child's anxiety
and parental limitations require that the child use the parents as
"anchor points." The child learns what it must do to engage the par
ents with a minimum of anxiety, and these modes of engagement
become the child's interactional template for all subsequent relation
ships. Out of these patterns the individual weaves the tapestry of the
self. The patterns learned in childhood are adhered to because they
were formed to avoid anxiety; if they are threatened, the individual
fears loss of contact, which is tantamount to loss of self and isolation.
According to Mitchell (1988), "each person is a specifically self
designed creation, styled to fit within a particular interpersonal con
text" (p. 277). If certain of these relational configurations are in conflict
with the predominant self-shaping relational patterns, they will not be
woven into the "dominant themes of the tapestry" and will find hid
den forms of expression, resulting in neurosis. Mitchell (1988) objects
to the classical concept of psychopathology that patients fail to meet a
standard of appropriate behavior and substitutes his view that "diffi
culties in living would be regarded with respect to the degree of
'adhesion' to one's early relational matrix and, conversely, the relative
degree of freedom for new experience which that fixity allows"
(pp. 277-278). It is the degree of rigidity of the relational configura

tions, that is, the extent of attachment to the archaic childhood objects,
that determines the extent of maladjustment of the personality.
Flexibility of the self-organization, the freedom to experience different
relationships in different ways, is Mitchell's concept of mental health.
As a direct consequence of this position, the aim of analysis,
according to Mitchell (1991), is to help the patient develop a more var
iegated sense of self. To achieve this goal, analysts have a twofold
task. First, they must immerse themselves in the experiential world of
the analysand, because it can be understood only from the inside.
Here, too, Mitchell is critical of object relations theories, as he believes
they attempt, much like the classical model, to understand the patient
from an "external" perspective. In Mitchell's view, the analyst must
inevitably become a part of the patient's relational world, and it is this
immersion that provides the understanding of who the patient is.
However, if analysts were to do only this much, they would be no dif
ferent from others in the patient's life. What allows for therapeutic
action, rather than simple repetition, is analytic inquiry. The analyst's
second task is to wonder with the analysand why his or her way of
relating is the only way he or she can form a relationship with the
analyst. According to Mitchell, the analysand was forced to choose
in childhood between his or her limited way of relating and total iso
lation. By finding neither inappropriateness nor incorrectness in
the patient's construction of the analytic relationship but, rather,
restrictiveness, the analyst creates the possibility of the patient's relin
quishing old object ties and becoming open to new interpersonal
experience. The aim is not to make the unconscious conscious, nor
even to provide a new experience, but to alter the structure of the
analysand's relational world.
In Mitchell's view, psychoanalysis as a treatment is a special type of
dyadic relationship that has as its aim the broadening of the relational
possibilities of one of the two participants. In the course of this
process the other participant, the analyst, will also be changed. But the
analyst's goal is to become both participant and observer of the
analysand's relational world so that he or she can offer a different per
spective on this world. The psychoanalytic encounter, then, creates the
possibility of broadening the structure of the analysand's relational
world beyond the narrow confines of childhood constraints.
A good example of Mitchell's concept of treatment is his approach
to narcissism. He takes a position midway between Kohut and Kern
berg. Mitchell (1988) views Kohut as understanding the childhood
need for narcissistic illusion but as missing the defensive function of
322 Chapter 7
grandiose and idealizing illusions in both childhood and in the adult

patient, and he views Kernberg as grasping the defensive nature of
narcissistic illusion but missing its function in normal development.
The child and the patient need both the illusion of narcissism and its
eventual relinquishment. In normal childhood the parent understands
the child's need for narcissistic illusion and engages in the child's play;
at the same time, the parent knows the illusions are play and must
eventually be given up. In Mitchell's view, the proper analytic stance
requires an analogous mixture of engagement in the patient's narcissis
tic fantasies, which Kohut understood, and analytic scrutiny of their
illusory status, which Kernberg understands. Mitchell advocates
engaging the patient and exploring why such illusions represent the
only way the patient can engage others. The combination of the ana
lyst's participation in the "play" of the analysand's narcissistic illusions
and his or her interpretations creates the possibility of experiencing
a relationship with other dimensions.
Mitchell accepts Sullivan's interpersonal theory of human experi
ence and his view of development as learning strategies of relating to
others. However, he adapts this model to a treatment strategy that
focuses on the patient-therapist relationship and aims to enrich the
patient's ways of interacting through this relationship.
Greenberg has recently become quite critical of some aspects of the
relational model he championed in the book he and Mitchell (1983)
wrote on object relations. Greenberg (1991) accepts the importance of
relationships but finds Mitchell and other relational theorists guilty of
committing three fundamental errors: (1) neglect of the fact that rela
tions are motivated by pre-experiential needs, (2) denial of the inher
ent nature of conflict in human existence, and (3) failure to recognize
the need for separation as well as relatedness.
Greenberg (1991) criticizes Mitchell, Fairbairn, and Kohut for fail
ing to appreciate that relatedness is not autonomous, but motivated
by other needs. He uses Fairbairn's and Kohut's theories as examples
of relational models that purport to view relations with others as
autonomously motivated but implicitly assume underlying drives that
fuel interpersonal contact. Fairbairn (1944) saw the infant as object
seeking but, according to his theory, object contact is necessary for
ego growth. Similarly, Kohut (1984) views the self-selfobject relation
ship as a means for the structuralization of the self. Greenberg argues
that both models assume an underlying motive for object contact
and therefore are implicitly drive models because relationships are
motivated and motivation is ultimately reducible to drives. Green
berg's conclusion is that no psychoanalytic theory can be sufficiently
explanatory without the concept of drives. Greenberg does agree with

Mitchell and other relational theorists that the biological concept of
drives is irrelevant to psychoanalysis, but he feels that these theorists
make the mistake of throwing the baby out with the bath water by
eliminating the drive concept entirely.
The second error Greenberg believes to be characteristic of rela
tional theories is their tendency to see human motivation as unidirec
tional. Examples are Kohut's view of development as moving toward
the structuralization of the self unless there is interference from an
unempathic environment and Winnicott's belief in the natural unfold
ing of the maturational process unless arrested by impingements.
Such theories, in Greenberg's view, deny the conflict inherent in the
human condition. In his view, conflict is not only produced by an
environmental counterforce but human motivation itself is inherently
conflictual; thus, tension is experienced in the most benign environ
ment. Greenberg believes that relational models tend to share the nar
rowness of the impulse/prohibition model of conflict by assuming
that all conflict involves defense. He argues that people frequently
experience conflicting motives without defense, that is, that the pres
ence of human conflict extends well beyond the employment of
defenses. Unidirectional relational theories are thus limited by their
inability to account for the ubiquity of human conflict.
The third limitation of relational theories, in Greenberg's view, is
their inability to appreciate the need to be separate and independent
of others. Greenberg argues that such a need is as critical to develop
ment as the need for relations, all autonomy is not reducible to
defense against object contact. Greenberg conceives of development as
an oscillation between relating and autonomy. In this regard, his
views bear a striking similarity to Mahler's (1975) separation-individ-
uation model discussed in chapter 1.
Greenberg's alternative to previous relational theories is a nonso-
matic drive/conflict model of motivation. He conceives of develop
ment as motivated by the dual needs for safety and effectance. Safety
encompasses the needs for physical and emotional well-being and
motivates relations with others. Effectance is the need to do and to
learn how to do; it provides a sense of vitality and aliveness to the
human experience but drives the individual away from others. Both
are endogenous, pre-experiential, nonbiological drives with innate
directedness. Both drives, in Greenberg's view, have immutable
underlying tendencies, but they are decisively influenced by interper
sonal events and therefore are expressed in a variety of behaviors.
According to Greenberg, there is inherent tension between the two
324 Chapter 7
drives. Safety drives people to others, and effectance requires doing

without others. The inborn nature of the two drives, therefore,
explains the ubiquity of conflict in the human experience. While con
flict may be caused by a negative response to the child's needs, such
responses are not required for the experience of conflict; the needs
themselves have opposing motivations. Greenberg illustrates this
point with an example of a child's hunger. The child may need the
hunger satisfied but may also feel the need to act autonomously. The
result cannot be fully satisfying: if the mother meets the need for
hunger she stifles the need for autonomy, and if she lets the child cry
she preserves his need for distance but leaves him hungry.
Greenberg believes his dual-drive theory preserves the importance
of relationships while resolving the three difficulties he finds in previ
ous relational theories. Nonetheless, a major developmental implica
tion of Greenberg's drive theory is a de-emphasis on the actual
child-mother exchange. According to Greenberg, many factors influ
ence the outcome of early experience, one of which is the relative
strength of the drives. Consequently, the meaning of any particular
developmental experience may result from perceptions of the parents
based on the child's needs and may have little to do with the actual
interaction between parent and child. Thus, Greenberg tends to weigh
internal motivation more heavily than do other relational theorists.
In Greenberg's view, the center of initiative of the personality is
best conceptualized as the ego, as Freud used the term, before the
structural model, when he referred to the "dominant mass of ideas."
Greenberg views the mind as an active container of ideas that cannot
be a direct object of experience but can be represented. Self-represen
tations and feeling states reciprocally influence each other and
are the primary regulators of behavior: "the way we are feeling at
any moment significantly affects how we imagine ourselves to be"
(Greenberg, 1991, p. 171). In turn, the shape of the self-image pro
foundly influences affective responses. Since self-representations are
multiply determined constructions, many ways of thinking of oneself
are possible at any given time. However, a particular self-representa
tion tends to be stable and dominant and decisively influences inten
tions, which in turn lead to wishes. Each wish consists of a self, an
object, and a relationship between the two. Whenever the dominant
self-representation is threatened, painful affects are experienced, stim
ulating repression to safeguard the integrity of the self. When devel
opment surpasses previous inabilities, the latter are "re-represented"
as negative. For example, when a child learns to read, its previous
inability to read is "re-represented" as shameful and repressed. All
defense, for Greenberg, is directed against such "re-representations."

According to Greenberg, the task of analysis is to understand the
meanings of representations by bringing the unconscious "re-repre
sentations" into awareness. The patient will not be easily amenable
to such broadened awareness because the repression of threatening
self-representations produces a feeling of safety. Resistance, accord
ing to Greenberg, is motivated by the safety drive. In this regard,
Greenberg's ideas are in agreement with the views of Weiss and
Sampson (discussed in chapter 1) that resistance is an index of the
lack of safety the patient feels with the analyst and that the analyst
must first help the patient feel safe before interpretations will be
accepted by the patient. Psychoanalysis, in this view, depends on the
patient's feeling of safety with the analyst, which in turn allows
rejected mental contents into awareness and thus broadens the
patient's self-representations. Since self-representations lead to
wishes, which involve both self- and object-representations, increasing
the flexibility of the self-representational field expands the patient's
motivations and allows for greater possibilities for effectance and
ways of interacting with others. Like Mitchell, Greenberg sees pathol
ogy as rigidity and effective analysis as a broadening of the interac
tional field, allowing the patient to respond to different interpersonal
situations in different ways.
Despite this similarity with Mitchell's thought, Greenberg differs
from Mitchell and other interpersonal theorists in the greater empha
sis he places on the patient's representational world. Because Green
berg believes that relational needs are half of the human motivational
system and are motivated internally, he is, despite his theoretical
divorce from the biological basis of psychoanalysis, closer to the
classical model than any other interpersonal theorist.
EDGAR LEVENSON
Another major extension of the Sullivanian model is proposed by

Levenson (1985) who, unlike Mitchell and Greenberg, accepts the
major tenets of Sullivan's interpersonal theory. In contrast to Mitchell,
Levenson sees a decisive difference between the interpersonal and
object relations viewpoints; in fact, Levenson (1989) has attacked
Mitchell's efforts to link the two models as an unwarranted ecu-
menicalism that threatens to subvert the interpersonal viewpoint
(Levenson, 1989). In Levenson's (1981) view, the crucial distinction is
not between the interpersonal and intrapsychic points of view but
between "reality behind appearance" and "reality in appearance."
326 Chapter 7
According to Levenson, the classical analytic model makes the error

of proposing that the patient's fantasies, dreams, and other pre
sumed manifestations of "the unconscious" are clues to the patient's
psychic reality and that the task of the analyst is to decipher the
meaning of symbols and appearances to uncover the "true" psychic
reality. As opposed to this view, Sullivan's interpersonal theory
sees the patient's difficulties as residing in interpersonal reality, the
nature of which must be delineated by the analyst from the relation
ship between patient and analyst. Thus, the patient's problems in
living are not an intrapsychic reality to be uncovered by decoding
symbols and disguises but distortions produced by interpersonal
anxiety experienced in the real world. Fantasies, in this view, are a
reaction to real interpersonal anxiety rather than motivators of
interpersonal perceptions. From the data of the analytic situation,
the analyst attempts to discern the pattern of the interpersonal rela
tionships causing the patient's problem. The aim of the analyst's
delineation of interpersonal patterns is not to demonstrate the
patient's projections from the past into the analytic setting but
to show that the very interpersonal patterns currently occurring
outside the analytic setting are also taking place within it.
In this context that Levenson takes issue with object relations the
ories. Because object relations theorists tend to seek to uncover the
patient's psychic reality, whether this be thought of as a "true self,"
a "nuclear self," or a "regressed ego," these theories, according to
Levenson, commit the same errors as does the drive model. That is,
they seek a "truer" reality under appearance, rather than attempt
ing to delineate the patient's interpersonal reality. For Levenson
(1985), the distinction between which comes first, fantasy or real
ity, marks the critical division among the different psychoanalytic
schools of thought.
One can see from this model that Levenson is attempting to adapt
the Sullivanian interpersonal viewpoint to the clinical method of
using the patient-therapist relationship as the primary analytic instru
ment. The Sullivanian model changes the principal analytic question
from "What does it mean?" to "What's going on around here?" (Leven
son, 1989). According to Levenson, the question is the same for object
relations theory, as it is for the drive model: both seek meaning. The
interpersonal analyst does not search for meaning but for understanding
the interaction between patient and therapist.
From Levenson's viewpoint, object relations theorists, no less than
classical analysts, ignore the reality of the patient's life. Like Sullivan,
Levenson seeks to know in detail the realities of the life before him. In
discussing a case of Silverman's, Levenson (1987) criticized the thera

pist for neither discussing nor inquiring into the details of the
patient's life outside of therapy. Unlike Sullivan, Levenson believes in
paying close attention to the patient-therapist relationship as an
example of the patient's interpersonal pattern and as the best lever for
effective change.
These considerations have a number of clinical implications. While
the patient-therapist relationship is the critical factor in treatment, the
process does not depend on the analyst's correct interpretation of it.
Indeed, in Levenson's view, the analyst's perception of reality has no
privileged status over the patient's. The two form an interpersonal
reality in which neither party has a special claim as arbiter. Further,
the analyst is important not as a fantasied or transference object but as
a real person with real qualities. According to Levenson (1982), the
analyst will inevitably become caught in the patient's interactional
pattern, but this participation must be authentic, "not merely sincere."
Every interpretation is a form of participation in the interaction,
enlarging not only the communicational matrix between the two par
ticipants but, simultaneously, the relationship per se. Levenson (1982)
breaks down the distinction between speech and action: all speech is
action, and all action is structured like a language. Thus, for Levenson,
psychoanalysis becomes a semiotic science. Everything that takes
place between the participants is a communication and is coded like a
language. According to this model, speech is only one form in which
such coded communication takes place. It is Levenson's contention
that everything that is talked about between the two participants is
simultaneously enacted between them. Consider, for example, a sce
nario in which the analyst points out that the patient is sensitive, and
the patient, upon hearing this, begins to cry. If the patient is masochis
tic, the analyst may feel benign and detached, that is, sadistic, or the
analyst may get angry at the patient and actually feel sadistic. In either
case, the sadomasochistic relationship is enacted (Levenson, 1982). In
opposition to the usual analytic posture, Levenson does not believe
that continual interpretation of this enactment will achieve an analytic
result. The therapeutic action, in his view, lies in "a changing, or at
least expanded, participation with the patient around the material. In
some way the therapist must operate with the patient so as to be
'heard'" (Levenson, 1982, p. 99).
Levenson (1991) acknowledges that this conceptualization of the
change process is obscure, but he believes that a certain degree of inef-
fability is inherent in therapeutic shifts. In Levenson's view, the
change process is stimulated not by information exchange but by
328 Chapter 7
interpersonal "resonance," which is difficult to define. Although the

precise nature of the process is unknown, the analyst delineates the
authentic patterns of the patient-analyst interaction in such a way that
he "resonates" with the patient's private experience and a patterning
emerges. Levenson's (1991) concept of resonance gives his view a sim
ilarity to the self-psychological concept of therapeutic action as affect
attunement (Kohut, 1984; Basch, 1984). Unlike self psychologists, how
ever, Levenson believes that when moments of resonance occur, the
patient experiences a sudden reconfiguration into which all the mater
ial seems to fit. Change, in this view, is discontinuous. Analytic tech
nique cannot cause this repatterning to occur nor predict when it will
happen; it can only attempt to stimulate the preconditions for its
occurrence.
Levenson views the analytic task as an enriching human relation
ship in which there is no place for the analyst to employ "techniques,"
such as interpretation, or "corrective emotional experience" from
outside the relationship. According to Levenson (1982):
Corrective emotional experiences largely disappear in the tar pit of the

patient's self-equilibrating system. I doubt that the patient grows because he is
supplied with a nurturing environment. I suspect the patient must be
engaged and experienced and responded to. If behavior is a language, then it
must be heard. The therapist who is detached from an angry patient may hear
him on the speech level but does not hear him on the action level [p. 100].
The patient is helped neither by a new truth, nor even by a new rela
tionship; rather, the patient's experiential world is enriched by the ana
lyst's authentic engagement in this world. In response, the patient gives
up his or her wish to change in favor of a wish to be his or her authentic
self. Psychoanalysis, from this viewpoint, is not a talking cure but an
"experience cure." The interpersonal analyst in this view does not nur
ture the patient or "accept the patient's psychic reality" but engages the
patient's world in a real way. One can see that Levenson's variant of the
interpersonal model stretches the concept of psychoanalysis beyond
both object relations theories and Mitchell's concept of the interpersonal
model. Authentic engagement replaces interpretation and the meeting
of childhood needs as the principal therapeutic instrument.
THE WORK OF MERTON GILL AND IRWIN Z. HOFFMAN
Another major extension of the interpersonal model has been pro

posed by Gill and Hoffman. Unlike Mitchell and Levenson, they do
not begin from a Sullivanian theoretical base, although Gill (1981) has
acknowledged the link between his "social model" and Sullivan's
interpersonal theory. Gill theorizes primarily about the analytic
process, with little reference to personality development or psy
chopathology. He contends that much analytic practice is of poor
quality because analysts do not pay sufficient attention to the
moment-to-moment here-and-now transference process owing to their
failure to see the patient's veiled allusions to the transference. In Gill's
view, Freud may have made this error himself since he did not believe
all of a neurosis would necessarily translate into the transference neu
rosis. The history of psychoanalysis is replete with failures to see that
seemingly nontransferential material is really a distorted allusion to
the transference. This problem is serious, in Gill's view, because the
extent to which the neurosis enters the transference is the extent to
which it can be resolved.
It follows that the analyst's first task is to expand the transference
within the analytic situation. This is accomplished by attempting to
link external material to the relationship with the analyst. Patients
talk about material external to the transference, which Gill considers
frequently to be a form of resistance, and analysts often resist con
necting this material to the transference because analysands may
attribute attitudes to their analysts that make them uncomfortable.
Nonetheless, analysis of resistance to the awareness of transference
is a crucial phase of the treatment process. Gill points out that the
analyst must be alert not only to the patient's attitudes but also to
the attitudes the patient implicitly (that is, nonverbally) attributes to
the analyst. Gill disputes the traditional technical stance of allowing
the transference to unfold spontaneously. His contention is that the
patient's resistance will prevent such an unfolding without interpre
tation of the resistance to awareness of the transference. The patient
will of course react to the analyst's persisting efforts to frame trans
ference interpretations, but the patient will likewise react to the ana
lyst's silence. Any reaction of the patient to the analyst is "grist for
the analytic mill" and should be interpreted.
The analyst must be careful not to increase resistance by telling the
patient he is "really" referring to the analyst; rather, interpretations
should emphasize the parallel between the external material and the
analytic situation. When such a parallel exists, the analytic reference
is the most relevant and has priority over all other material, including
genetic interpretations, from an interpretive point of view. This prin
ciple applies at any point in treatment. The analyst need not wait for
the development of the "therapeutic alliance," because transference
330 Chapter 7
interpretations will help foster that relationship and without such

interpretations the patient's resistance may prevent its formation.
Once the patient is aware of the transference, Gill believes the cru
cial issue is to demonstrate to patients that their way of construing
the analytic situation is not the only way. He disagrees with the tradi
tional view that the patient is "distorting" the relationship, a view
that reflects what he considers to be the positivistic classical view of
the analytic situation, namely, that it has an objectively perceivable
reality that the analyst knows and that the patient must come to see.
In this regard, Gill's epistemological position is in agreement with
that of both Mitchell and Levenson. In Gill's view, patients do not
distort but they do not see what they are bringing to the relationship;
Gill terms this "resistance to the resolution of the transference." This
resistance is overcome when patients see that their version of the
relationship is their own construction, which is not, however plausi
ble, the only possible way of viewing the analyst's behavior and the
analytic relationship in general.
Up to this point, Gill's views appear to apply only to the conduct of
analysis. What shifts Gill's (1981) position from a purely technical
to a theoretical model is his view that the transference is always
connected to the real behavior of the analyst. In every analysis the
analyst evinces both technical and personal behavior, and the patient
responds to both. Gill believes there are major clinical implications in
the view that the analyst's behavior, both technical and personal, is
real and is responded to by the patient as both transference and real
ity. Because the transference is always connected to the real analytic
situation, reference to it should be included in all transference inter
pretations. By pointing out what in the analyst's behavior has con
tributed to the patient's perceptions, the interpretation becomes more
plausible and acceptable to the patient. Furthermore, if the analyst
cannot find the connection between the transference perception and
his or her actual behavior, the task is to uncover it.
One can see in Gill's view a shift from the typical analytic conceptu
alization of transference as the patient's projection of wishes and fan
tasies onto the analyst to an interpersonal model according to which
the patient is always reacting to a real person and the transference is
seen as an interactional phenomenon. Object relations theories and the
drive-ego model tend to hold similar structural views of transference:
it is what the patient projects from his past onto or into the analyst
and the analytic situation. Gill, like the Sullivanians, opposes this con
cept of transference in favor of the interpersonal view that the real
relationship between patient and analyst is inevitably implicated in
the transference and its interpretation. The clinical implication is that

the actual features of the analytic situation must be clearly delineated
and even scrutinized. Gill (1981) sums up his view of the therapeutic
action of psychoanalysis this way:
The analyst suggests that the patient's conclusions are not unequivo
cally determined by the real situation. Indeed, seeing the issue in this
way rather than as a "distortion" helps prevent the error of assuming
some absolute external reality of which the "true" knowledge must be
gained. The analyst need claim only that the situation is subject to vari
ous interpretations and that since the patient's conclusions are not
unequivocally determined by the features of the situation which can be
specified, he would be wise to investigate how his interpretation may in
part be influenced by what he has brought to the situation [p. 118].
There is an even deeper level at which Gill's theory is interpersonal.

For Gill (1983), the fact that the analytic relationship is interpersonal
means that the analyst will inevitably "fall in" with the patient's trans
ference wishes to some degree. Therefore, the meaning of the analyst's
participation must be explored in every analysis. This investigation
does not necessarily imply "countertransference confession," although
such confession is not precluded in every situation. An exploration of
the relationship between patient and therapist in which the participa
tion of both is reflected on becomes a crucial component of the analy
sis. Like Mitchell and Levenson, Gill views the analyst as an inevitable
participant in the patient's pathology, however, in opposition to those
theorists, Gill believes that interpretation is the primary tool for the
resolution of the transference neurosis. Gill agrees with the Sulli
vanians that the analyst is a participant but believes in the therapeutic
efficacy of both the new experience and its interpretation. Gill's view
is that analysis works best when what is being experienced between
patient and analyst is simultaneously interpreted.
Hoffman (1991), who has collaborated with Gill on a study of trans
ference, has emphasized the epistemological aspects of their shared
viewpoint, which he believes leads to a "social-constructivist" model
of the analytic process. He believes that the division in psychoanaly
sis is not so much between the drive and relational models as between
the "positivist" and "constructivist" positions. He describes the posi
tivist model in the same way Gill does: the myth of an objective reality
knowable in some privileged fashion by the analyst. Like Mitchell and
Levenson, Hoffman opposes such a view and believes, instead that the
patient and analyst are coparticipants in an interactional drama in
332 Chapter 7
which together they are continually constructing an interpersonal

reality. Every act by either party, such as an interpretation, shapes the
interaction in a new, not totally foreseeable way. Each participant has
a perspective on the "social construction," but neither has a privileged
view of it. Hoffman agrees with the interpersonal emphasis of much
of object relations and interpersonal theory, but he contends that these
views do not follow through and embrace the proper epistemological
perspective. He attacks such theories as latently positivistic because
they imply that one can explore the patient's perceptions to "get at
something already there" (Hoffman, 1991).
The alternative offered by Hoffman's social-constructivist model
is for the analyst to recognize that any exploration may lead to
something never before formulated that may affect both partici
pants in unforeseeable ways. The explication of the patient-analyst
relationship and the influence of each participant on the other cre
ate new meanings. In Hoffman's view, the adoption of this model
implies a shift in the analyst's attitude toward a recognition that
intervention cannot capture a "reality" and therefore shapes the
process in unknowable ways. In Hoffman's (1991) view, this atti
tude leaves analysts freer to engage patients in a more open,
authentic manner because they are not burdened with the myth of
believing in a "right" interpretation that they must find but have,
instead, the perspective that they cannot foresee the results but are
ready to explore them. In Hoffman's view, this sort of analytic
encounter differs from a simple, existential confrontation because
such analysts continually reflect on themselves and their involve
ment. The aim of this process is "to affect some of the most deeply
rooted ways in which patients experience themselves and others"
(Hoffman, 1991, p. 96). Further, the patient is made aware that these
patterns are relative rather than absolute and inevitable, and this
constructivist attitude will be carried over to the rest of the patient's
life.
SUMMARY OF THE INTERPERSONALISTS
One can see from this review that the interpersonalists differ in many
ways but are united in the view that man is defined by his relation
ships with other people. With the possible exception of Greenberg,
interpersonal theorists believe that the inherent nature of personal
relationships is neither incidental nor even simply necessary for the
achievement of human aims but the very substance of life. Gill and
Levenson both invoke this principle as the key difference between the
interpersonal and object relations models. All versions of the latter

model see interpersonal relationships as necessary for the formation
of psychological structure whereas the interpersonalists see man's
very nature as inherently relational. Gill (1983) points out that for
object relations theories pathological issues occur in the context of
interpersonal relations whereas for interpersonal theories issues are a
matter of interpersonal relations. If one views the drive-ego model as
relatively neglectful of interpersonal relations and the object relations
model as emphasizing them, the interpersonal model is the next step: a
theory based on interpersonal relations.
In reviewing the interpersonal theories one is struck by the rela
tively minimal attention given to concepts of development and
pathology. Sullivan is the only interpersonal theorist who had a
detailed theory of development. Since Sullivan, interpersonal theory
has tended to focus on man-in-relationships of various sorts, with rel
atively little attention to the phases and issues in development leading
to their occurrence. One can see in the work of Levenson, Gill, and
Hoffman a relative neglect of the developmental viewpoint in favor of
the view that relational problems exist in the present and must be
addressed as they manifest themselves in treatment. Mitchell gives
more consideration to the influence of development on current prob
lems but still focuses his treatment approach on the enrichment of cur
rent relationships in the patient's life. Greenberg is the exception, as
he believes that the interpretation of "re-representation" of the past is
crucial in helping to free the patient from rigid self-representations.
Since Sullivan did not draw on his developmental theory in his treat
ment approach, Greenberg is the only major interpersonal theorist
whose clinical strategy includes a significant role for interpretation of
childhood experience.
Conceptualizations of psychopathology also tend to be absent
from the post-Sullivanian interpersonalist theories. Sullivan himself
addressed specific diagnostic categories from his interpersonalist per
spective, but all major interpersonal theories since have concentrated
on concrete interpersonal difficulties in living, not clinical syndromes.
This focus is a direct product of the interpersonal principle that people
are defined by their interpersonal relationships. Since this principle
is no less applicable to people who seek help than to others, interper
sonalist therapists tend to regard clusters of symptoms or character
structure as being of little clinical relevance in comparison to the inter
personal strategies patients use to negotiate their lives. Common to
all versions of the interpersonal paradigm is the idea that as these
strategies are altered, existing difficulties will be simultaneously
334 Chapter 7
ameliorated. The "diagnosis," rather than invoking a traditional nosol

ogy, is framed in terms of interpersonal patterns, and such patterns are
believed to be best seen in the patient's relationship with the analyst.
This emphasis on the interpersonal nature of human problems
has led theorists in the post-Sullivan movement to place the patient-
analyst relationship at the center of the analytic process, just as those
who follow object relations theories or classical psychoanalytic techni
cal theory do. However, there is a major difference in the way this
relationship is construed. The interpersonalists locate the analyst's
participation in the patient's issues at the center of the analytic pro
cess. Levenson (1981) even states that analysis begins when the
participants experience what is talked about. Gill (1981) has quite cor
rectly pointed out that this view of the analyst as participant sepa
rates the interpersonal theory of technique from both the classical
and object relations models.
The interpersonal view leads logically to the question of the relative
importance of experience versus interpretation in therapy. Theorists
differ on this issue. Sullivan, of course, gave little role to interpretation
(Havens, 1976). Levenson (1989, 1991) has concluded that authentic
encounter is more conducive to change than is the accretion of knowl
edge or even the provision of a new relationship. Hoffman (1991) sees
a role for interpretation but emphasizes authentic encounter. Mitchell
(1988) believes in the importance of interpretation but not because it
offers mutative insight; for Mitchell, the new relationship with the
analyst is the most significant therapeutic factor and interpretations
help consolidate this relationship by showing the patient "where the
analyst stands." Again, Greenberg (1991) is traditional in his belief
that interpretation is the center of the therapeutic action. Gill (1981)
also gives interpretation the most significant role in the analytic
process; in fact, Gill insists, much like a classical analyst, that every
thing that goes on ought to be interpreted as fully as possible. That is,
although Gill, like all interpersonal theorists, sees a major role for the
analytic relationship in effecting change, his focus is clearly on inter
pretation: the therapeutic relationship is optimally formed by consis
tent transference interpretation, and the new relationship is most
effective when it is interpreted.
The exclusive focus on the interaction between patient and analyst
has led interpersonal theorists to mount epistemological challenges to
psychoanalytic theory. Since analyst and patient bring their respective
backgrounds to a conjointly shaped relationship, the privileged episte
mological position of the analyst is abandoned in favor of the view
that neither participant has privileged access to the truth about what
happens between them. In this sense, the interpersonal view is linked

to certain extensions of self psychology, such as the intersubjective
approach of Stolorow and his colleagues (1988) and Schwaber's (1983)
view of analytic listening. According to all these theorists, the analyst
is not the arbiter of reality but a coparticipant with a separate, fre
quently differing, viewpoint. Consequently, in Levenson's version of
interpersonal theory the analyst offers not the truth, but only his or her
experience of the situation. Gill and Hoffman, by contrast, hold to the
primacy of interpretation but emphasize that the aim of the interpre
tive process is to open the patient to alternative perceptions. The epis-
temological issues raised by the interpersonalists are not abstract; they
deeply influence the posture of the practicing analyst. Levenson and
Hoffman both emphasize these implications of interpersonal thought.
Both interpersonal and object relations viewpoints shift psycho
analysis from a theory of intrapsychic wishes and defenses to a theory
of development and pathology connected inevitably to interpersonal
relationships. However, one can see from this review that despite the
similarities in theoretical perspective between the interpersonal and
the object relations/paradigms, there are clear distinctions between
them that render the two models fundamentally different. First, in
object relations theories interpersonal relationships are a means to an
end, whether this end is conceived of as the formation of the self or as
ego growth; in all object relations theories, personal relationships lead
to psychological structure, and this structure defines the health or
pathology of the personality. Each variant of interpersonal theory, on
the other hand, views the interpersonal relationships themselves as
actually constitutive of personality with little concept of psychological
structure. This difference leads to distinct views of the analytic situa
tion. In interpersonal theory the analyst is a co-participant in a rela
tionship whereas in object relations theories, the relationship is
primarily a construction of the patient, and the analyst's role is to
understand it and sometimes meet the needs expressed through it.
Gill believes that this difference—the analyst as a co-participant and
the analyst as an interpreter of the patient's relational experience—is
the key cleavage in modern psychoanalytic theory.
Second, all object relations theories include a theory of develop
ment and a concept of the relationship between developmental diffi
culties and subsequent types of pathology. One does not find such
correlations in interpersonal theory; they are obviated by the empha
sis on current interpersonal patterns. For object relations theories,
developmental concepts are crucial because the analysis aims to
achieve its goals by making contact with a part of the personality that
336 Chapter 7
was either arrested or inhibited by conflict at some point in develop

ment. Current problems are seen as having direct connection with
developmental issues. In each variant of the interpersonal paradigm,
with the exception of Greenberg's formulation, developmental issues
are either largely ignored (for example, Levenson, 1985, and Gill, 1981)
or related to current problems in a complex manner that does not have
significant treatment implications (Mitchell, 1991). Greenberg's (1991)
emphasis on re-representation in the interpretive process gives the past
the most prominent role among interpersonal theories.
Third, the epistemological issue divides the object relations and
interpersonal paradigms. Levenson (1989) and Hoffman (1991) believe
that the division in psychoanalysis is between the "positivist" view
point of the analyst knowing "objective reality" and the view of real
ity as relative to the observer. Whatever epistemological position the
various object relations theorists may adopt on the nature of an objec
tive reality, they all see the role of the analyst as one of understanding
the patient—whether that understanding is used for interpretation,
the meeting of needs, or the formation of a new relationship. By con
trast, the interpersonalists tend to view the analyst's role in terms of
presenting a point of view and offering a potentially enriching rela
tionship. Gill and Hoffman emphasize the analyst's role as interpreter
more than other interpersonalists, but they, especially Hoffman, stress
the relativity of the analyst's perspective. Thus, the goal of interper
sonal psychoanalysis is to overcome problems by expanding, trans
forming, and enriching the nature of the analysand's world of human
relationships; the aim of object relations analysis is to change the
structure of the personality.
CRITIQUE OF THE INTERPERSONAL SCHOOL
The interpersonal viewpoint in psychoanalysis adds a significant

dimension to the efforts of object relations theories to shift psychoana
lytic theory and practice from an impersonal drive model toward a
theory of the person relating to objects. As is clear from this review,
the interpersonal theories, while sharing the interpersonal conception
of personality and analytic process, form a widely divergent, often
conflicting, group. They all point out that however much object rela
tions theorists may purport to shift the concept of the analytic process
from the drive model to a notion of the person in relation to others,
they all retain a tendency to view the analyst solely as an interpreter
of the patient's experience. Even Winnicott, who tended to view the
analyst's role as active adaptation to the patient's needs, employed
interpretation in the traditional manner of interpreting the patient's

experience. By contrast, the interpersonal position is that the analyst
is inherently a participant in the very process he or she seeks to
understand.
Nonetheless, some of the later contributors to object relations
theorizing have tended to emphasize the interactional role of the
analyst. The tendency toward an interactional model in Kleinian
thought, as seen most prominently in Racker's work, Stolorow's
extension of self psychology to an "intersubjective" approach, and
Green's application of Winnicott's thought are all examples of a
movement in object relations theories toward a more interperson-
ally oriented view of the analytic process. Although these develop
ments are a major connecting point between object relations and
interpersonal theories, there is still a decisive difference between the
two models. For the interpersonal school, the assessment and treat
ment of personality are a matter of interpersonal relationships. The
object relations paradigm tends to view relationships as the primary
influence on the development of psychological structure rather than
as a substitute for it. Consequently, object relations theories tend to
emphasize the role of early relationships in the formation of stable
internalized perceptions of self and others whereas the interper
sonal theorists either disregard development (Levenson, Gill and
Hoffman) or view it as a series of relationships (Sullivan, Mitchell,
Greenberg). In the conceptualization of the treatment process, the
interpersonalists place far greater emphasis on the analyst's role in
shaping the interaction than do even those theorists who propose
an interactional extension of object relations theories.
Mitchell adds a great many specific conceptualizations to the gen
eral contributions of the interpersonal paradigm. He points out the
significance of the relational matrix in development and contempo
rary problems in living. More than most other interpersonal theorists,
he emphasizes the ubiquity and importance of conflict, thus blending
interpersonal theory with the conflict model. Furthermore, Mitchell
underscores the active role of the child in forming attachment to its
parents, thus combining the environmental dimension of interper
sonal problems with the child's contribution. Finally, Mitchell has
made an invaluable contribution with his explicit recognition of the
critical human need to adhere to the established sense of self and of
the intense anxiety attendant upon threats to the self structure. While
most object relations theorists recognize the importance of self struc
ture, Mitchell has specifically emphasized the patient's investment in
his or her sense of self as an element blocking analytic change.
338 Chapter 7
Despite these contributions, there are some fundamental draw

backs to the several versions of interpersonal theory. First, the episte-
mological position that reality is relative and that the analyst's view is
no closer to the truth than any other tends to be contradicted in the
work of each theorist who advocates it. Levenson (1989) contends that
the analyst has no special claim to know the patient's reality but sees
the task of analysis to be the exposing of the patient's distortions of
interpersonal reality. Hoffman (1990) has pointed out that Levenson
advocates a perspectivist position and then proceeds to treat his own
interpretations as observations. For example, Levenson discusses a
dream of one of Kohut's patients in which the patient plunges a knife
into a straw doll and draws blood. Kohut interpreted that the patient
was disappointed that the analyst was not the strong father he had
hoped for. Levenson (1989), wishing to go beyond Kohut's interpreta
tion, asks if the dream could not mean "exactly what it says, namely,
that the therapist is real, that he can be hurt, that he is vulnerable" (p.
549). Hoffman points out that although Levenson's interpretation is
"cogent," it is only an interpretation (whereas Levenson himself
regards it as "exactly what [the dream] says"). Hoffman finds this
objectifying tendency to be a general problem with Levenson's
approach. Further, Hoffman points out that Levenson tends to regard
the patient's view as reality, and that in the dream example, the
patient, according to Levenson, sees the "reality" that the analyst is ill
and vulnerable and the analyst has no claim to a truer reality. This
position reverses the classical view that the analyst is "right" and the
patient "wrong" to a comparably positivistic stance in which the
patient is "right" and the analyst "wrong."
Levenson (1990) replied to Hoffman that he is not a perspectivist
and that he in fact erred in implying that he was. Levenson disagrees
with Hoffman's view that reality is ineffable and that one can only
have a plausible perspective on it. For Levenson, there is a reality, but
it is obscured by distortions in the patient's life due to interpersonal
anxiety and the analyst has no special claim to know it. He sees his
own formulations as no "truer" but as efforts to grasp the "real experi
ence of the patient." As Levenson (1990) states, "The point is not that I
believe that my perceptions are direct observations of interpersonal
events, but that my observations focus attention on interpersonal
events, not the patient's distortions" (p. 302). Levenson (1990) goes on
to attack Gill and Hoffman's position that patients must see that there
are plausible views of the analytic situation other than their own. He
believes that this position, like Stolorow's, fails to take the final step of
accepting the patient's perceptions not as a "point of view" but as a
"valid basis for inquiry into real events in the patient's life and in the
analysis" (p. 301). In Levenson's view, Gill and Hoffman "straddle the
fence of reality" by accepting the patient's view as "plausible" but
insisting that the patient see other views. According to Levenson, this
position actually requires patients to see that their insistence on their
point of view is wrong; analysts may not tell patients that they are
wrong, but they communicate this very judgment by suggesting that
there are other ways of looking at the situation. For Levenson (1990,
1991) this type of treatment is a form of persuasion.
One can see that in his attempt to defend himself against Hoffman's
charges Levenson makes his position less clear. While insisting that
the analyst has no privileged claim to know reality he says that the
patient's problem is the distortion of reality and that the analyst's com
ments focus on interpersonal events, "not the patient's distortions." If
the analyst has no claim to reality, how can he discern distortions and
how can he have access to "real" interpersonal events? If he does not
have "direct observations of interpersonal events" to offer, how can he
make observations "on interpersonal events"? Levenson's position
results in the conundrum that the analyst's task is to focus on interper
sonal reality even though the analyst cannot claim to know it.
Despite this confusion in Levenson's epistemological position, he
makes a valid criticism of Gill and Hoffman. Levenson (1989) points
out that Gill and Hoffman claim to disavow any notion of a superior
analytic view of reality but proceed to describe the transference as
"not a distortion, but a misreading" (p. 302). Levenson contends that
the type of analysis advocated by Gill and Hoffman is based on the
following assumptions: there are right and wrong views for the
patient to adopt; if the patient views the analyst in one particular way,
he or she is "resisting," and this perception must be worked on until
the patient sees that there are other views; the analyst's view that
there are other "plausible" interpretations is clearly the view that the
patient is enjoined to adopt. Hoffman (1991) even says that the patient
has to become a "constructivist." For Levenson, such a prejudgment of
analytic outcome raises the question of subtle persuasion.
Further, the very usage of the concept of resistance suggests an
inconsistency in Gill's and Hoffman's epistemological position. There
is no element more embedded in the psychoanalytic concept of the
analyst as the arbiter of the patient's reality than the notion of resis
tance. Indeed, some theorists who are closer to the classical model
than Gill have abandoned usage of resistance because of its judgmen
tal connotation (for example, Giovacchini, 1979). However one may
attempt to soften it, resistance means that the analyst knows that the
340 Chapter 7
patient is excluding something from his or her awareness, even if the

analyst does not know what it is. If the patient disputes the analyst's
interpretation, he is "resisting." Hence, the analyst is the arbiter of
reality. One cannot meaningfully speak of the patient as "resisting"
except in this sense of the analyst's judgment that the patient is
excluding from awareness realities about himself.
To be sure, Gill and Hoffman attempt to address this problem by
insisting that the patient's perception is plausible and that what the
patient resists is not some "truer" way of experiencing but the mere
consideration of other plausible ways. However, this rejoinder does
not entirely solve the problem. First, it applies only to what Gill terms
"resistance to the resolution of the transference," not to "resistance to
the awareness of the transference." That is, the patient's failure to see
an allusion to the transference in what he or she is saying or experi
encing is itself deemed a form of resistance. The patient must be
brought to accept the reality that the material is at minimum also a
reference to the transference. And the analyst is arbiter of this reality.
Moreover, even in the context of "resistance to the resolution of the
transference" the patient's refusal to adopt the analyst's view that there
are other interpretations of the analyst's behavior is another example of
how the patient fails to see "reality." Gill and Hoffman refer to the
patient as resisting not any particular interpretation but the comple
mentary "truths" of other ways of looking at the same situation.
According to Gill and Hoffman, if patients persist in their positivistic
commitment to their own "truth," they are resisting and the analyst's
view, rooted in his or her superior constructivistic epistemology, may
be taken to be correct.
There is still another inconsistency in the perspectivism advocated
by Gill and Hoffman. While they insist that analysts can never know
the validity of their interpretations, they also point out that some
interpretations are better than others (Gill, 1991; Hoffman, 1991). The
problem is that they make this claim while eliminating any epistemo-
logical basis for judging any interpretation as better than any other. If
they believe that one interpretation is better than another, they must
provide criteria by which "better" and "worse" can be assessed, crite
ria that then become the basis for the view of reality posited by the
analyst and offered to the patient.
The same problem applies to Hoffman's (1991) effort to differenti
ate his view of analysis from existential encounter by citing the ana
lyst's continual reflection on his or her participation in the process.
Any criterion for such a reflection assumes a view of reality by which
this process may be judged. Here one arrives at the fundamental prob
lem with the constructivist view of the analytic process: either the con
structivist must hold that there are no criteria for judging any particu
lar interpretation as better than any other, in which case there can
truly be no analysis, only one individual offering his or her experience
to another in an existential encounter. Or, the constructivist must
admit, as Gill and Hoffman have, that some interpretations are better
than others, in which case they implicitly invoke, without admitting it,
some criteria for making such a judgment, a position that is difficult to
differentiate from the ordinary analytic view.
The reason Gill and Hoffman are given to these inconsistencies is
that they are trapped in the conundrum of attempting to implement
their constructivist philosophy without eliminating the very basis for
analysis. Since they do not wish to reduce analysis to existential
encounter, they are ultimately forced to utilize implicitly the very
notions of reality and validity of interpretation that conflict with their
constructivism. We found the same problem with Levenson's con
tention that while the analyst has no access to "reality," the analytic
task is to focus on the patient's distortions of it. The interpersonal the
orists have argued persuasively that analysis is inherently an interper
sonal enterprise, but they have not been able as yet to successfully
sustain the view that the interaction of two people with equally valid
viewpoints is legitimate analytic treatment. Since each version of
interpersonal theory ultimately wishes to defend the analytic model,
each theory resorts to the very concepts of reality and validity of inter
pretation it purports to dispute. Interpersonal theory has yet to
resolve this epistemological dilemma.
The epistemological position of interpersonal theory leads to the
relative depreciation of the concept of psychopathology. Patients do
not "have" an objectively ascertainable form of pathology that they
bring to the analysis. This position fails to account for just why the
patient has emotional problems. The concept of a pathological person
ality seeking relief conflicts with the interpersonalist notion that
patients' problems inhere in their relationships. Since interpersonal
theories tend to minimize the concept of psychopathology, they must
develop an alternative view of why emotional distress occurs.
Mitchell's contributions are limited by this problem. Mitchell (1988)
contends that there is no normative human personality against which
people can be judged. Consequently, pathology is not a deviation but
an extreme degree of "adhesion" to early relational patterns. As we
have seen, Mitchell views maladjustment as rigidity of experience.
However, this criterion does not hold up to scrutiny. Pathological pat
terns cause inherent difficulties in relating to others, leading to
342 Chapter 7
depression, anxiety, and distress, irrespective of their degree of flexi

bility. For example, borderline patients are far more pathological
than neurotic or "normal" individuals not because their patterns are
more rigid but because of the types of demands they make on oth
ers and the d ifficulties they have in functioning with others
(Summers, 1988). Such patients become lonely and distraught not
because they have only one way of relating but because that way
does not allow them to form satisfactory bonds with others. If their
one way of relating were more amenable to the formation of gratify
ing relationships, they would be far happier, even if lim ited.
Neurotics may be equally rigid, but since their relational patterns
do not cause as much difficulty in forming bonds, they are less dis
turbed. In short, Mitchell appears to confuse lack of resilience in
personality patterns with maladjustment.
Mitchell's reluctance to acknowledge that some needs are pathologi
cal per se is illustrated by his view of patients with dependency needs.
Such patients are not pathologically dependent, in Mitchell's view;
these needs are "perfectly appropriate," as all people are dependent.
The problem, as Mitchell sees it, is that the patient has excessive anxiety
over normal dependency needs. Such a position is contradicted by the
nature of the expectations and demands of many patients. For example,
consider borderline patients, who make all manner of demands on the
analyst—such as to hold and caress them, understand and gratify wants
never expressed, and give them money and material goods—for imme
diate gratification of needs. The fact that such patients fall into despair if
their needs are not met immediately underscores that these needs are
not "perfectly appropriate" (Summers, 1988).
There are also difficulties with Mitchell's concept of neurosis as
"loose threads" that do not fit into the tapestry of the personality. First,
in neurotic character pathology, a healthy aspect of the personality is a
"loose thread." Moreover, many neurotic patterns are woven well into
the personality structure; as Mitchell himself points out, patients have
great difficulty yielding these patterns precisely because they fit so well
into the sense of self. This fact points out the contradiction in Mitchell's
two criteria: if maladjustment is tantamount to rigidity of self experi
ence, it is difficult to see how neurosis can be a "loose thread" that does
not fit into the sense of self. Indeed, if neurosis really does not fit in, it
should be easily amenable to change because it would not be part of the
self that is the cause of pathological rigidity. In sum, Mitchell has made
an impressive effort to explain "difficulties in living" without resorting
to a judgment of psychopathology, but his two criteria are in conflict
and neither can explain emotional problems satisfactorily.
The same problem of equating rigidity with pathology is inherent

in Gill and Hoffman's variant of interpersonal theory. They offer no
explicit developmental theory to account for the formation of emo
tional problems, but since their treatment focus is on inducing the
patient to recognize other ways of perceiving the analyst, their con
cept of treatment is equivalent to a loosening of rigid interpersonal
ways of perceiving. Their views thus fall prey to the same conflation
of singularity of experience with pathological patterns. In fact, some
patients acknowledge that there are other ways of viewing the ana
lyst's behavior but still cling to their original view, which may be, for
example, that the analyst does not like them. In these circumstances, it
is the tendency to view people as hostile or assaultive that approaches
the crux of the problem, not the rigidity with which this attitude is
held. Even more tellingly, there are many characterologically dis
turbed patients who are inconsistent and even chaotic in their percep
tions of the therapist. For such patients, a major component of the
pathology is severe and unpredictable transference oscillations: the
therapist is now an angel, later a devil; now a savior, later a persecutor
(for example, Kernberg 1976, 1984; Summers, 1988). The pathology of
such patients is best captured by the lack of correspondence of the
perception to reality, not its inflexibility. Because Gill and Hoffman
are reluctant to use the concept of reality in their explanations of
pathology, they are confined to the criterion of rigidity.
Thus the problem of explaining emotional problems is unresolved
in the interpersonal paradigm. Having abandoned the notion of psy
chopathology in favor of the concept of a person relating to other per
sons in a particular way, the interpersonalists are left to explain
without making any judgments about pathology why some people do
less well than others. The result is the attempt to employ concepts
such as "rigidity" or "inability to fit into the personality structure" as
explanations, but these issues can neither explain the differences
among types of pathology nor differentiate adequately between
pathological and healthy adaptation.
All of this explains why Mitchell (1988) and Levenson (1991) tend
to view psychoanalysis not as a treatment but as an "enriching experi
ence." This conclusion follows logically from their view that people's
problems inhere in relationships and that these relationships may not
be judged as reflecting personality disturbance per se. The point of
analysis in such a view is the provision not so much of a new experi
ence as of an "enriching" relationship that widens the patient's experi
ence. This goal is not only unduly modest for many patients who seek
and need more than enrichment, but also inappropriate for certain
344 Chapter 7
patients. Many depressed and anxious patients can have enriching

relationships and still remain depressed and anxious. An enriching
relationship would not reorganize the structure of a narcissistically
disturbed patient unless such a relationship helped to facilitate
understanding of his needs for idealization and grandiosity.
In sum, enriching relationships do not necessarily resolve the fun
damental object relations conflicts that produce symptoms. For
Levenson (1981), for example, the aim of pointing out the patient's
interactions with the analyst is to connect these interactions with the
patient's interpersonal patterns outside the analysis. It is difficult to
see how the awareness of this connection would resolve the interac
tional pattern. Both Levenson and Mitchell have a limited concept of
how pathology is actually resolved, and their view of analysis there
fore seems quite restricted. Gill and Hoffman have a more elaborated
concept of therapeutic change, but it is difficult to see how adoption
by patients of a constructivist perspective necessarily resolves their
problems; the acknowledgment of alternative possibilities cannot be
expected to lead to the growth of the self, any more often than do
other forms of "enriching experience."
The limited view of analysis advocated by the interpersonal school
is in sharp contrast with the more ambitious aims of the object rela
tions theorists. Because the latter do embrace the concept that pathol
ogy can be grasped by the analyst, they tend to believe that pathological
patterns can be illuminated, worked through, and resolved. We have
seen that the various object relations theories share the traditional
analytic assumption of psychopathology residing in the patient but
depart from classical analysis in the belief that pathology consists of
object relations units. Apart from these presuppositions, object rela
tions theories differ and even conflict among themselves. This diver
sity returns us to the question of whether one can discern a common
object relations model of the analytic process that can serve as an
alternative to the drive/conflict model. Having elucidated the main
principles of the various interpersonal theories and differentiated
them from object relations theories, we are now in a position to
address this issue.
CHAPTER 8
An Object Relations Paradigm

For Psychoanalysis
The m a jo r d if f e r e n c e s a m o n g o b je c t r e l a t io n s t h e o r ie s w o u l d se e m
to militate against the concept of a unified object relations model.

Nonetheless, these very real differences should not obfuscate the
commonality of principles that differentiate an object relations
approach from both the drive model and the interpersonal model.
By a critical elucidation of principles one can delineate an object
relations paradigm within which the various theories can fit as
variations. Such a model cannot be constructed by an uncritical
eclecticism nor by a preference for one particular theory to the
exclusion of the others. The approach taken here is to build on the
critical assessments of each theory to define an object relations con
cept of psychoanalytic theory and practice that fits no particular
theory but illuminates object relations as a model differentiated
from the drive-ego and interpersonal models.
The common principle of all object relations theories is the view
that the fundamental human motivation is for object contact rather
than drive discharge. As we have seen, both Klein and Kernberg
adhere to the drive theory but view the importance of drives to be not
so much in their ability to discharge tension as in their role in the for
mation of the object relations that become the building blocks of the
psyche. It may be said that all object relations theories view the forma
tion of object relations as the primary human motivation. Further, the
concept of drives in both the Klein and Kernberg views does not hold
up to scrutiny. Klein's view of drives is the same as Freud's: they are
innate, biologically driven impulses for tension reduction. This con
cept does not fit the modern concept of drive as a hierarchical organi
zation that is responsive to stimuli and able to adapt to changing
circumstances (Tinbergen, 1951; Lorenz, 1963). Kernberg's (1976)
notion of drives, unlike Klein's, is informed by modern conceptual
izations. However, since in Kernberg's view a drive can only be
expressed through an object relationship, the drive concept is obvi
ated, especially since aggression, which is central to Kernberg's
345
346 Chapter 8
theory, does not fit the drive concept. When one eliminates the unnec
essary concept of drives, the motivational concept in the theories of
Klein and Kernberg corresponds to the concept in the theories of
Fairbairn, Guntrip, Winnicott, and Kohut, namely, that the organism
is inherently directed toward object contact.
PSYCHOLOGICAL STRUCTURE
Any object relations theory leads to a concept of the self, the devel
opment of which is linked to the vicissitudes of the object relations
units. As Andre Green (1977) has pointed out, the complement of the
object is not the ego but the self (see chapter 4). As we saw in chapter
2, Fairbairn (1944) used the word ego, but employed it to refer to the
self, not the ego of the tripartite model. Since development is not a
matter of the taming of drives, in object relations theories the impor
tance of ego structure wanes in favor of the development of the self
structure, which is a product of the internalization of attachments in
the form of object relationships. In this way the object relations para
digm replaces the development of id and ego with self and object rela
tions as the cornerstones of development. The way the self structure is
experienced at any given moment is the sense of self; that is, the phe
nomenological experience of the self, the sense of self, is a reflection of
the underlying self structure.
While it is implicit in most object relations theories that psycho
logical structure is built with object relations units, Kernberg (1984)
makes this point explicit and uses the concept of object relations
structure to understand pathology. Furthermore, he is alone in point
ing out that the psychological structure of the neurotic is composed
of object relations units. However, for Kernberg the structure of
object relationships is the structure of the ego. Given the data of child
development, to be described shortly, suggesting that attachments
rather than drives are primary and given the fact that the drive con
cept is unnecessary even in his theory, it makes more sense to con
ceive of the structure built from object relations units as self structure
rather than ego structure.
The data from infant research do not support the concept of a sepa
rate id (White, 1963; Stern, 1985). There is a compelling quality to
some of the infant's behavior, but this "drivenness" applies as much
to searching for stimuli as to biological gratification (Lichtenberg,
1983). From the beginning of life the neonate will interrupt feeding to
look at stimuli, but there is no need to postulate "independent ego
energies," as the developmental data do not support giving primacy
An Object Relations Paradigm 347
to drive gratifications. As we have seen, the evidence indicates that

the infant is best conceived as programmed to seek contact and rela
tionships of various types, only one of which involves the satisfaction
of biological urges.
This view of psychological structure as self formed from object rela
tionships leads to a different view of anxiety than that proffered by
the drive-ego model. According to the latter, anxiety is a warning
signal to the ego of threat from id or superego (Freud, 1926). While
Freud did acknowledge that traumatic anxiety, the feeling of infantile
helplessness, is characteristic of early infancy, nontraumatic anxiety
encompasses most forms of anxiety. From the object relations view
point, anxiety is a threat to the sense of self, to one's sense of who one
is. Consequently, object relations theories tend to describe anxiety as a
threat to existence rather than as the eruption of particular contents
into consciousness. This reconceptualization links Klein's "annihila
tion anxiety," Winnicott's "unthinkable anxiety," and Kohut's "disin
tegration anxiety," all of which refer to a threat to one's psychological
existence. These terms were not used solely to describe severe pathol
ogy but were deemed relevant to development and pathology of
almost any type. Mitchell (1988) and Greenberg (1991), from the view
point of interpersonal theory, also emphasize the threat to self (or self
representation in the case of Greenberg) in the experience of anxiety.
Since, from the object relations viewpoint the self structure is com
posed of object relations units, any threat to the structure of those
units is a threat to the self and will be experienced as anxiety.
Therefore, the self may be threatened even in well-integrated person
alities. This notion of anxiety also fits Sullivan's concept of anxiety as
a threat to the self-system.
According to the drive-ego model, anxiety is caused by a threat of
eruption of a particular content, such as a childhood wish eliciting
guilt or shame, and defended against by a particular defense, such
as repression. The problem with this view is that if the repressed
content is made conscious, guilt or shame is felt; it is difficult to see
why a guilt-invoking wish should evoke a sense of threat. Guilt is
not a pleasant emotion, but it causes feelings of "badness" rather
than threat. As we saw in chapter 2, this reasoning led Fairbairn and
Guntrip to conclude that shame, rather than guilt, is at the source
of most pathology. However, while shame is often a more painful,
threatening feeling than guilt, it does not necessarily evoke anxiety.
In fact, neither shame nor guilt can account for the threatening
experience of anxiety. In the object relations paradigm, anxiety is a
threat to one's sense of who one is, and this explains the sense
348 Chapter 8
of dread, or angst, even in the well-integrated neurotic individual.

The type of anxiety emphasized in object relations theories is closer
to Freud's concept of "traumatic anxiety." However, Freud (1926) con
signed this type of anxiety to the helplessness of the young infant
whose ego is so weak that he or she becomes overwhelmed. This psy-
choeconomic concept defines anxiety in terms of quantity of stimuli in
relation to the strength of the ego. The annihilation anxiety of object
relations theories is most usefully considered a threat to the sense of
self, the anchor point of one's existence.
DEVELOPMENT
The object relations view that attachment is autonomously motivated

fits the research data on development much more closely than does
the drive model. Three separate lines of research evidence support the
view that the infant is inherently motivated to make contact with its
caretaker rather than discharge tension. First, it seems that infants are
programmed for human contact and are preadapted to form a rela
tionship with the caretaker. Findings from experimental research
show that the neonate turns to the source of human sounds, differenti
ates human from nonhuman sounds, reacts with distress if the sound
source is disengaged from the view of the speaker's mouth, recognizes
the human face, and behaves differently to human and nonhuman
objects (Lichtenberg, 1983; Stern, 1985). In the first few weeks of life
the infant's feeding behavior becomes regulated to the particular
behavior of its caretaker. According to Stern (1985), the infant has an
active "social" life from birth. Beebe, Jaffe, and Lachmann (1992) have
demonstrated that the mother-child dyad develops its own mutually
regulated rules of communication, which are not predictable from
either partner separately. They conclude that the interactional
sequence between both parties to the mother-child relationship is
internalized by the child rather than by the mother as an object. Beebe
and Lachmann (1992) provide convincing evidence that in the first
months of life mother and child learn to match their affective direction
and interpersonal timing in a variety of ways and that the child comes
to expect this matching and reacts negatively if it is absent. Their
data provide convincing evidence for the view that a presymbolic
representational world consists of these mother-child interactions.
Evidence against the view of infant behavior as tension reducing
comes from the fact that infants pursue objects visually from the
beginning weeks of life and will even interrupt feeding to look at visu
ally presented objects (White, 1963). The findings that infants actually
seek stimuli and make active choices for preferred stimuli with no
consequent tension reduction is strong evidence against the concept of
the infant as a discharge-seeking organism (White, 1963; Stern, 1985).
Infants will react aversively to noxious stimuli only. The view that
emerges from the experimental data is of an infant programmed
almost from birth to form a synchronous interaction with its caretaker
and to regulate stimulation both by increasing and decreasing it
(Lichtenberg, 1983).
The second primary source of evidence for the object relations view
of motivation comes from naturalistic observation and experimental
work with animals. John Bowlby (1969) marshaled an impressive
array of ethological evidence demonstrating the existence of a power
ful need for attachment among nonhuman primates. Newborn guinea
pigs, lambs, and dogs will attach to physical objects or animals of
other species if those are the only possibilities for contact without
receiving other sources of gratification. Indeed, puppies isolated for
three weeks and punished upon their only contact with a human will
become more attached to that figure than will puppies receiving
rewards for human contact. Equally compelling are Harlow's experi
ments with rhesus monkeys, which showed that baby monkeys attach
to cloth model mothers and not wire models even when latter provide
bottle feeding (Harlow and Zimmermann, 1959). Further, rhesus mon
keys raised by a nonfeeding cloth mother surrogate will attach to the
model mother and cling to it when alarmed or in a strange setting
(Harlow, 1961). However, baby monkeys raised by a wire model that
provides bottle feeding do not use the surrogate for comfort despite
strange or dangerous conditions. Like lambs and puppies, monkeys
cling intensely to the cloth mother in the face of danger and punish
ment, even if the danger is from the surrogate mother itself. Bowlby's
evidence shows that nonhuman primates have an autonomous need
for attachment irrespective of gratification of biological needs and that
this attachment endures.
Bowlby's (1969) work also summarizes the third line of evidence:
naturalistic data on children showing that they attach to figures who
do not meet their physiological needs. Children in concentration
camps reared without the opportunity to attach to a benign adult
formed strong bonds with each other. A study of Scottish children
showed that about one-fifth attached to adult figures who did not par
ticipate in their physiological care. Furthermore, Bowlby pointed out
that there is as yet no evidence that human babies attach to adult fig
ures because of their association with the meeting of biological needs.
Bowlby (1969) summarized the evidence this way: "Such evidence
350 Chapter 8
as there is strongly supports the view that attachment behavior in

humans can develop, as it can in other species, without the traditional
rewards of food and warmth" (p. 218). This work extends the concept
of the autonomous need for attachment to humans.
Ethological and child research evidence provides solid support
for the object relations view that there is an autonomous human
need to attach to significant figures very early and that this attach
ment tends to endure. These findings fit with Fairbairn's view of the
infant as inherently object seeking rather than pleasure seeking.
According to the experimental data, infants are preadapted to seek
object contact and to form an interactional synchrony with the care
giver. The object relations view adds to the experimental data the
concept that the human, being a symbolic animal, will at some point
in development give meaning to these attachments and use them to
develop a sense of self. According to this object relations model, the
structure of the self is formed from the internalization of early
attachment relationships and is based on the symbolic meaning the
child gives to its early object ties. In this way these early object rela
tionships not only endure but define the sense of self and influence
later relationships with others.
Each object relations theory has a different way of understanding
the process by which early object contact forms the sense of self. For
Fairbairn and Guntrip, the ability to love without destruction and the
degree of schizoid withdrawal resulting from fear of the destructive
ness of object contact are crucial. For Klein and Kernberg the vicissi
tudes of aggression and its integration with good objects are the
critical components in the formation of the self. For Winnicott, the
availability of the "environmental mother" to provide background
support for the natural unfolding of the self is central to the internal
ization of the good mother, which allows the maturational process
to unfold. In Kohut's view, the gradual relinquishment of archaic
grandiosity and idealization, fostered by the mixture of parental grati
fication and frustration, results in the structuralization of the self. Each
object relations theory has a different concept of how the self develops
and of the role the environment plays in its formation. While one can
not justifiably speak of a single object relations theory, each theoretical
viewpoint is based on the general principle that autonomously moti
vated attachments to early figures become internalized and form the
building blocks of the self.
In the object relations view of development, children navigate
the interpersonal world by forming object attachments designed to
minimize anxiety and pain. However, they will always prefer some
attachment, no matter how painful it may be, to loss of contact. Mitchell

(1988) has pointed out that children form the type of relationship they
need to maintain contact with early objects and that these relationships
play a crucial role in the formation of the self structure. As indicated,
experimental evidence indicates that young subhuman primates will
attach to animals who only cause pain if no other alternative is avail
able (Bowlby, 1969); in fact, they will attach more strongly than will
young animals who are not abused. According to the object relations
paradigm, children prefer painful attachments to isolation in order to
avoid the anxiety of annihilation. This principle is the object relations
explanation for the observation that children and adults form intense,
resilient bonds with harmful figures. Fairbairn pointed out that Freud's
drive theory resorted to the "death instinct" to explain this phenome
non, an explanation that is exceedingly speculative and has few backers
even within the classical tradition. In the object relations view, children
form bonds to their caretaker early, the bond with an abusive adult
tending to be even stronger than that with a benign adult. Further, early
relationships are internalized to form a self structure that is then
defended against all threats. Attachments to early negative figures and
their persistence in childhood are explained by the need for bonding, as
is the persistence of the effects of such attachments into adult life.
As Mitchell (1988) has pointed out, these internalized object rela
tionships are not simply passive imprints of the relationships; they are
the ways in which children construe the relationships and what they
extract from them according to their defensive and adaptational
needs. However, Mitchell (1988) is incorrect in attributing a passive
view of the infant to object relations developmental arrest theories.
Winnicott (1963a, b) emphasized the activity of even the very young
infant in "creating" the object; while initially this "object creation" is
purely fantastical, in later infancy and childhood the creation of the
object out of "transitional space" plays a primary role in development.
Similarly, Fairbairn emphasized the child's active object seeking
in personality formation and pathology. While Kohut's views lend
themselves to the interpretation of the child as a passive recipient
of parental empathy or injury, he did see the child as defensively
responding to parental failures.
All these theories take into account the child's response to disrup
tive environmental events and suggest that the way the child con
strues and responds to the environment plays a critical role in the
formation of psychological structure. For example, if the parent is abu
sive, the child may be too traumatized to internalize the parent as
attacking and sadistic, and may defend against this internalization by
352 Chapter 8
a defensive construction of the parent as benign; that is, the child

internalizes an attacking, sadistic parent layered over by a defensive
"benign" parent. Each such constructed internalization is a building
block in the edifice of psychological structure. This process tends to
become complex: object relations units that defend against others may
conflict or harmonize with each other, and still other object relations
units may be used to defend against such conflicts. The buildup of
these object relations units in their variegated patterns forms the
unique structure of each individual self.
Kohut (1971, 1977) emphasized more than any other theorist the
construction of the self structure from the child's internalization of
aspects of the parenting figures in both normal and pathological
development. The emphasis in Kohut's theory was on the infantile
self as bipolar, consisting of the poles of grandiosity and idealiza
tion, each of which must be supported by parental gratifications.
With each experience in which the infant relinquishes some degree
of its archaic narcissism, there is accretion of psychological struc
ture. In Kohut's view, the key to emotional development is the for
mation of self structure. The data from infant research support
Kohut's view that the gaze of the parent is an important aspect of
development (Lichtenberg, 1983; Stern, 1985). At birth the infant sees
best at eight inches, the right distance for breast- or bottle-feeding.
Further, if the infants do not successfully elicit the parental gaze, they
actively search for it, and if they cannot find it they will eventually
withdraw. One may therefore conceive of the child as programmed
for "mirroring" interaction with the caretaker. These findings support
the importance of mirroring in development, though they do not sup
port Kohut's emphasis on the mother as the sole initiator of the mir
roring relationship. Kohut's view ignores the fact that the infant is
primed for mirroring, and proceeds actively to seek it, rather than
just being in need of it. Fairbairn's conceptualization of the infant as
object seeking and Mitchell's (1988) view of the infant as actively
seeking and forming relationships are better supported by the data
from infant research.
The object relations view is that development hinges on the abil
ity of the parent-child interaction to foster the growth-enhancing
object relationships out of which a functional self is constituted.
Winnicott and Kohut both recognized the importance of parental
empathy and support in the development of the self, but Kohut
underemphasized the child's role in the construction of the mother-
child relationship. Infant research has underlined this emphasis
by demonstrating that the child actively seeks out a synchronous
parent-child interaction and by documenting the aversive response

of the child to not receiving it (Lichtenberg, 1983).
PSYCHOPATHOLOGY
If the parent-child attachment does not foster growth-enhancing

object relationships, the formation of the self will be arrested and its
functioning crippled. For example, if the dominant attachment is to a
denigrating parental figure, the child may well defend against the
internalization of an object that devalues the self by adopting a supe
rior, aloof posture to the now denigrated object. In this case, a signifi
cant portion of the self structure consists of "object devalues self" and
"self devalues object." These object relations structures will interfere
with efforts to achieve satisfying interpersonal relationships. In this
situation the self consists of dysfunctional object relationships that
impede effective functioning.
Each object relations theory has a different view of how crippling
object relationships come about. For Fairbairn (1940) and Guntrip
(1969), the mother's inability to accept the child's love leads to the
internalization of "desirable deserters," against which the child must
defend itself. For Klein and Kernberg, if the internalization of bad
objects exceeds that of good objects, pathological processes such as
splitting and projective identification ensue. In Winnicott's view, if the
environment is unable to provide for the child's needs within the dif
ferent phases of dependence, the internalization of the good mother
is thwarted and the development of the self is arrested. According
to Kohut, if the mother is unempathic with the child's needs for
grandiosity and idealization, transmuting internalization is blocked,
archaic grandiosity and/or idealization remain, and the self is not
able to develop. In every variant of the object relations paradigm,
psychopathology is a product of distorted object relationships that
interfere with the structure and functioning of the self.
In this object relations paradigm, Freud's criteria of mental health—
to love and work well—are assumed. The difference between this par
adigm and Freud's conception lies in the reasons for the inability to
operate well in these two arenas. In each variant of the object relations
view, the integration and smooth functioning of the self allows for
achievement and interpersonal satisfaction. Some versions of interper
sonal theory also fit this view of adjustment. Mitchell's (1988) concept
of psychological health as a well-integrated self formed from rela
tional patterns is compatible with the object relations view. However,
as pointed out in chapter 7, Mitchell's concept of pathology as rigidity
354 Chapter 8
of relational patterns and his concept of neurosis as a "loose thread"

in the tapestry of the self are questionable and do not accord
with the object relations root of neurotic symptoms. For example,
Mitchell (1988) refers to Stern's view that because issues recycle
throughout life, one cannot assume that a particular issue originates
in a specific developmental phase. Mitchell concludes that issues
frequently judged to be pathological, such as the dependence-
autonomy conflict, are normal and that the patient's needs are "per
fectly appropriate." This view misses the fact that it is not the exis
tence of the conflict between dependency and autonomy that makes
a personality pathological but the meaning of each need embedded
in an unproductive object relations structure that makes depen
dence and individuation so difficult to achieve. When the object
relations structure does not allow the meeting of one or both of
these needs, the functioning of the self is impaired.
Fairbairn viewed all neurosis as a manifestation of the conflict
between dependence and autonomy. However, he recognized that
what makes an individual symptomatic is not the existence of the con
flict, which is ubiquitous, but the inability of the individual to form a
satisfactory object relationship with a separate object. Object contact
threatens the self, and the individual withdraws. The conflict between
the need for object contact and autonomy is not the source of the
pathology; the neurosis arises from the fear of object contact, a fear
that causes difficulties both for the attainment of autonomy and for
the establishment of relationships. The following principle is recog
nized in all object relations theories: while any individual may be
caught in conflicts between competing needs or desires, what makes
such conflicts pathological for some is an underlying object relations
structure that interferes with the achievement of satisfaction in one or
both of the conflicting needs. In a similar fashion, Greenberg (1991)
sees the fundamental human conflict as a battle between the need for
safety, which motivates relationships, and effectance, the need for
individuation. Greenberg's view of the human dilemma as a balanc
ing act between these conflicting drives bears a striking similarity to
Fairbairn's concept of conflict between the needs for dependence and
autonomy. The difference is that for Greenberg the conflict itself is the
cause of the pathology, with maladaptive relational patterns resulting
from the conflicting needs.
While it is true that every object relationship limits the self in some
way, the pathological individual has had to accept a malfunctioning
self in order to achieve object attachment. Fairbairn and Guntrip see
the result of the compromise in all pathology as an inability to relate
to objects in an authentic manner. According to Klein and Kernberg,

because the integration of aggression into the self is compromised,
pathological mechanisms are used to preserve the self. For Winnicott,
this compromise is viewed in terms of arrested maturation of the self.
For Kohut, the price paid is a realistic sense of self-esteem that allows
the achievement of healthy goals and ambitions.
An illustration of pathological self formation from the viewpoint
of self psychology is Kohut's case of Miss F. As discussed in chapter
6, in analytic sessions this patient tolerated neither extended periods
of silence nor any of Kohut's comments that went beyond a mere
summary of her material. When Kohut attempted the latter, Miss
F became enraged and felt the analysis was being ruined. Kohut
reached the conclusion that these reactions were not resistance, but
the emergence of the patient's grandiose self and the parallel need for
mirroring. Through the development of a relationship in which she
was only echoed, Miss F experienced the sense of strength and
wholeness of self that was missing from her childhood. Because the
only object contact of which she was capable consisted of the satisfac
tion of her need for mirroring, in order to achieve a sense of self she
needed to experience the bolstering of her grandiose self by a selfob
ject. That is, her self was only minimally functional since it was able
to operate only within the narrow parameters of a reliable selfobject
who performed the mirroring function. Miss F's narcissistic pathol
ogy consisted of this severe impairment of self functioning and the
consequent limitation of her goals and ambitions.
Guntrip's analysis of the schizoid character illustrates the extreme
of compromise to maintain a minimal sense of self. As we saw in
chapter 2, Guntrip's formulation was that schizoids are so hungry for
love that they fear destroying the object with their desires for merger.
Since all affect threatens to stimulate the hunger for merger, the
patient removes affect from all interactions and assumes a defensive
aloof posture in order to protect the object and the self. By this charac
ter defense of withdrawal the patient is able to maintain contact but at
the cost of emotional poverty, emptiness, loneliness and apathy.
A case example of this schizoid compromise is Guntrip's patient,
discussed in chapter 2, a young woman who worked through her
traumatic rejection by her grandfather in her therapy. When her fear
of trusting others was linked by Guntrip to her relationship with him,
she began to withdraw from him. Guntrip then interpreted that she
had been more trusting of him lately, and she responded by becoming
cold and uncommunicative and said that she feared she could never
respond to another human being. This cold, hardened attitude was
356 Chapter 8
the only way she could relate to others without feeling the terror of
affective object contact.
Both Kohut's and Guntrip's cases demonstrate that even the inter
nalization of painful object relations units helps to achieve a sense of
self and allay anxiety. In some situations, the child is forced to inter
nalize painful and even destructive object relationships in order to
avoid anxiety and achieve a sense of self connected to others. The
child will do so because the internalization is better than the alterna
tive—annihilation anxiety. Analogously, all internalizations, no matter
how seemingly healthy, have a cost because they restrict interac
tional patterns in the very process of anxiety avoidance. Thus, every
accretion of self structure through object relationships has both an
adaptive value and a limiting effect on the functioning of the self.
Pathology and health are a question of the balance between the
restrictive, dysfunctional aspects of object relations units and their
benefits. The extent to which internalized object relationships inhibit
the development and functioning of the self is the extent of pathology
in the personality.
This view of pathology questions the conventional view, held even
by some object relations theorists, that object relations theories apply
to early, or preoedipal, issues and the drive-ego model applies to neu
rotic conditions, presumed to originate in the oedipal phase. Mitchell
(1988) is critical of "developmental tilt" theorists who maintain the
classical view that conflict between drives and defenses is the primary
issue at the oedipal stage and who insert object relations "under
neath" as the key to earlier development. This effort to accommodate
object relations and drive theories characterizes Winnicott's (1963a)
and Kernberg's (1984) view of the neuroses, as well as Kohut's (1971)
early work. On the other hand, Fairbairn (1941), Guntrip (1969), Klein
(1937), and Kohut (1977, 1984) in his later work all viewed the rela
tionship between object and self as critical to all pathology, including
neurosis. As Mitchell (1988) has pointed out, if development is a mat
ter of the relationship between self and object, drives have no greater
import in later stages than in earlier ones.
In the object relations paradigm proposed here, psychopathology is
tantamount to an edifice built with dysfunctional object relations
building blocks. This concept of psychopathology can accommodate
the classical view of neurosis as conflict between wish and defense
only if wish and defense are both conceived of as object relations
units. This solution has been adopted by Kernberg (1984), who consid
ers neurosis to be a result of psychic conflict and the elements of the
conflict to be object relationships. We saw in chapter 5 that Kernberg's
view of the inherent nature of object relationships from the earliest

phase of psychological life obviates the drive concept and defines neu
rosis in terms of conflicting object relations units. As we saw in chap
ter 5, Kernberg's clinical illustrations of analysis for neurotic conditions
were difficult to distinguish from his treatment approach to character
pathology because in both cases split object relations units were ana
lyzed and then synthesized. In no case was Kernberg able to establish
that defense against a wish resulted in a neurotic symptom.
If the self structure consists of object relations units, neurosis must
be a matter of dysfunctional object relationships (Summers, 1993b).
Even if the neurotic symptoms are manifestations of conflicts between
object relationships, the units themselves must be defective in some
manner to account for a symptomatic outcome. One is then left with
the question of differentiating neurosis from other forms of psy
chopathology. On the basis of object relations theory, the criterion for
this differentiation is the extent of dysfunction resulting from the
object relations structure of the personality. In the borderline person
ality, severe impairment in the abilities to love and work results from
the need to fuse with the other to complete the sense of self (Summers,
1988). The object relationships of fusion and hostile defiance to protect
fragile self boundaries are sufficient to achieve a minimal sense of
existence but leave little opportunity for effective functioning
(Kernberg, 1975; Summers, 1988). The higher level characterologically
disturbed patients gain more functional capacity from their object
relations but are crippled in relationships, work, or both. Patients are
labeled neurotic because their object relationships allow functioning in
major areas of their lives but impede success and gratification.
Severity of pathology is thus a matter of degree: the extent to which
the object relations structure of the personality interferes with the
functioning of the self is the degree of psychopathology to which an
individual is subject.
In the object relations paradigm, then, neurotic pathology reflects a
defect in the structure of the self. This view stands in opposition to
Mitchell's metaphor of neurosis as a "loose thread" in the "tapestry"
of the self, a notion that reflects the classical view of neurosis as symp
tomatic outcome in an integrated personality. But the loose threads
metaphor does not adequately address the pain, suffering, and func
tional impairment of the neurotic individual. The self dysfunction typ
ically labeled neurosis is more like a tear that weakens the fabric
without ruining it completely. Thus, the aims of analytic treatment for
the neurotic involve restructuring the self rather than reintegrating
loosened components (Summers, 1993b). This view bears a kinship to
358 Chapter 8
Kohut's (1984) later concept that neurosis consists of a cohesive sense

of self that has been unable to achieve its life program. This notion
fits closely the concept of neurosis as an adaptive object relations
structure that impedes effective functioning. However, the underlying
object relations structure that impedes the achievement of the "nu
clear program" of the self is far more complex and variegated than is
suggested by the concept of transmuting internalization of grandiosity
and idealization. As noted, Stolorow, Brandchaft, and Atwood (1987)
adopt a multimodal concept of self that moves self psychology closer
to viewing psychopathology as a product of a complex object relations
structure.
The mixture of adaptive functioning with pathological constraints
explains the ubiquity of conflict in neurosis without putting it at the
source. The functioning of the self is continually obstructed by patho
logical object relationships that pressure the self in a different direc
tion. Neurotic conflicts result from these pulls in opposite directions;
however, the pathology is not caused by the conflict itself but by the
maladaptive object relations units. Clearly, object relations theories
tend to de-emphasize the function of guilt in pathogenesis in compari
son to the classical model, but they do consider the role of excessive
guilt in neurotic conditions. The crucial question for the object relations
paradigm is whether guilt interferes with the functioning of the self.
To the extent that it does, it can be a pathogenic factor in neurosis—as
well as in more severe pathology.
The reason guilt is so relatively underemphasized in object rela
tions theories is that guilt is not frequently a significant element of
crippling object relations structures. Recall that Fairbairn (1944) and
Guntrip (1969) viewed guilt as a defense that provides the illusion of
being in control when one in fact feels weak and ashamed. Kohut
(1984) made a point of his belief that conceptualization of human
dysfunction has shifted from Guilty Man, who denounces his own
immorality, to Tragic Man, who is defective, weak, and incapable.
Green (1977) expressed a similar belief in his statement that psycho
analysis has changed its paradigmatic figure from Oedipus to Hamlet.
These views represent the shift to an object relations view in which
pathology is rooted in dysfunctional self structure that produces feel
ings of shame and incapability, rather than in unacceptable wishes
that give rise to guilt.
To be sure, the theories of Klein (1937), Winnicott (1960a), and
Kernberg (1976) accord a more central role to guilt in pathology, espe
cially neurosis. For these theorists, guilt originates in the anxiety of
realizing one can injure the mother one loves and depends on. In this
view, excessive guilt consists of disabling object relations. Thus, in the

Kleinian concept of the depressive position and in Winnicott's emen
dation of it as the stage of concern, guilt is an outgrowth of anxiety
rather than of superego strictures.
While Kernberg (1976,1984) does seem to accept the classical view
of neurotic guilt as oedipally formed and organized, he acknowledges
that even in the neuroses the structure of the ego consists of units of
object relations that must be modified and reorganized. Ultimately,
even in object relations theories that include a potentially pathogenic
role for guilt, the concept does not fit the drive-ego model of a struc
tured superego in conflict with unacceptable wishes. When it is a fac
tor in pathology, guilt is viewed as another form of defective object
relationship inhibiting self functioning.
This conception of neurosis is illustrated by the case of a 40-year-old
real estate developer who, despite a seemingly successful long
standing marriage and some professional success, entered analysis
because he felt that something was wrong with his life and that
he was a professional "underachiever." It turned out that he had
been close to bankruptcy due to a business venture that was extremely
undercapitalized and that he had several other tumultuous and unpro
ductive business associations. Early in the analysis the patient com
plained, almost without noticing it, of most people in his life. The
employees upon whom he most depended tended to be either highly
volatile, unpredictable, and crisis prone or unresponsive and incom
petent. Consequently, his business life was in constant turmoil, and he
felt continually frustrated. His current business was making enough
money to provide a comfortable living, but he was in continual debt
and his success was far more limited than would have been expected
given his situation. As a result, his life was beset with continual tension
and anger.
Without entering into the details of a complex analysis, it may be
pointed out that the patient's business difficulties were traced to a
self-defeating pattern that manifested itself in a variety of ways, such
as by not acquiring sufficient capital, failing to receive sufficient funds
for costs, underselling his services, and hiring and keeping inadequate
and difficult employees. This pattern was related to the internalization
of his father as a mean, abusive alcoholic whom he feared but longed
to please. Every move toward success meant a threat to his internal
ized father, who would be envious and depreciative of the son who
longed to please him.
The object relations structure of the patient's bond with his father
consisted of a longing to idealize the father as strong, a fear of his
360 Chapter 8
father as abusive, and the intense desire to please his father and gain
his approval, which the patient defended against with anger. He was
in fact deeply identified with his father, and his business career began
with his entrance into his father's business. His father was a highly
successful businessman, and the patient's identification with him led
to a certain degree of business achievement. However, he was sub
jected to constant verbal abuse for trivial errors, and every successful
step led to fear that his father would be threatened and would with
draw approval and depreciate him. The patient's attempts to be like
his father and his fear that he would become an "abusive drunk" led
to a lifetime of oscillation between success and defeat. To be unsuc
cessful would have displeased the father and subjected the son to
vituperation, so the patient internalized the object relations unit of the
"good boy" trying to gain approval from the successful father.
However, every step of success threatened the internalized, envious
father, thereby stimulating identification with the "abusive drunk" he
did not wish to be. Consequently, the patient would achieve a certain
degree of success and then sabotage his ventures. This pattern was on
external reflection of his internalized ambivalent relationship with the
father who both wished him success and was threatened by it.
This abbreviated formulation of one aspect of this analysis appears
to fit the classical view of neurosis as unconscious conflict. The patient
was most certainly in conflict between two object relations patterns
involving himself and his father. However, the pathological pattern
was not a product of the conflict between the internalized "good"
father who wished him success and his fear of the threatened father
who wished to see him struggling and ultimately defeated. It is this
latter object relations unit, being pathological per se, that creates func
tional difficulty. The source of this patient's neurosis was not to be
found in the conflict between the pathological and healthy compo
nents of the self but in the very existence of the pathological compo
nents. This case illustrates the concept of neurosis advanced here.
The patient had a father-son object relations unit that allowed and
even encouraged functioning—until the internalized father was
threatened, at which point the son's professional success was inhib
ited. This layered father-son object relations structure fostered some
success but did not allow for expansive professional or financial
achievement; every time the patient made money, he found a way to
lose it. The aspect of the father-son dyad that led him to please and
identify with a successful father was highly adaptive and encouraged
success, but the object relations unit that made him fear the abusive,
denigrating father resulted in his self-defeating life pattern. The
patient's external life of oscillating successes and defeats thus mir

rored his internal world of productive and defeating object relation
ships. This structure corresponds to an adaptively neurotic pattern of
functioning that falls short of realizing the potential of the self. Within
this structure, neurotic conflict results from the coexistence of adap
tive and maladaptive aspects of the personality; the conflict is not the
cause of the unhealthy components of the self but their result.
This object relations paradigm of psychopathology has profound
implications for psychoanalytic treatment, and it is to the clinical
consequences that we now turn.
TREATMENT
It follows from the object relations paradigm of psychopathology that

the aim of treatment is to change the structure of the patient's object
relationships in such a manner that the self can function more effec
tively. The target of object relations analysis is neither the wish nor the
interpersonal relationship but the object relationship. We have seen
that most object relations theorists tend to depart from the classical
model by questioning the exclusive role of interpretation. There is
wide variation in the extent to which other interventions are advo
cated and interpretation is construed as offering a therapeutic experi
ence as opposed to new knowledge. Only Klein and Kernberg have
held to the classical concept that insight alone has therapeutic efficacy.
But even Klein viewed the mutative factor as the internalization of
the good object, and her followers (see Rosenfeld, 1978; Segal, 1981)
have emphasized the patient-therapist relationship as a crucial thera
peutic ingredient. In fact, a dominant theme in object relations theo
ries is the role of the analytic relationship in effecting therapeutic
outcome. Consequently, in the object relations paradigm there are two
categories of intervention relevant to the aims of psychoanalysis:
interpretation and the provision of a new relationship.
The Interpretive Model
The most direct and significant implication of the object relations par
adigm for psychoanalytic interpretation is that interpretations are cast
in terms of a self in relation to an object.
Analysts frame their interpretations not in terms of what the
patient wishes for or desires but in terms of the self-affect-object unit,
as Kernberg (1976) points out. Moreover, the elucidation of one such
unit is only the beginning. Because each object relations unit is a
362 Chapter 8
component of the self structure, the eventual target of the analytic

process is not a group of disparate units but the structure that they
form and the sense of self to which they give rise. The patient is never
angry without the anger directed at someone, and the self-angry-at-
the-other is embedded within a structure of other such units. For
example, the anger may be a defense against other, more threatening,
object relationships, such as longing for or fear of the other; or the
anger may be a means of seeking vengeance for the frustration of yet
another object relationship, such as a grandiose self-superior-to-
others. We may say that in this paradigm the analyst's task is to eluci
date the object relations units, including the defenses against these
units of experience, as they are enacted and reenacted with the analyst.
Since interpretations based on this paradigm are directed to a per
son in relation to another, the analyst points out to the patient what he
or she is doing rather than that he or she has a wish. This implication
is consistent with Schafer's (1976) notion of analytic interpretation as
"action language." Eschewing the concept of psychic mechanisms,
object relations analysts do not view the patient as enacting or subject
to psychical forces but as a person navigating relationships with
other people. Fairbairn (1943) referred to this type of intervention as
the "personalization" of interpretations and pointed out that inter
preting according to the drive model fosters defenses by distancing
the patient from the material. This principle may well be adhered to
by analysts who do not identify themselves with object relations theo
ries, but insofar as this is the case such analysts are departing from the
classical psychoanalytic paradigm with its emphasis on drives.
The most immediate and accurate path to the patient's object rela
tions structure is the patient-analyst relationship. From the object rela
tions perspective, transference is the perception of and interaction
with the analyst according to the object relationships formed from the
patient's early attachments. With the exception of Sullivan's approach,
interpretations are geared as much as possible to the transference
within all models of psychoanalytic treatment, but the object relations
paradigm has a stronger theoretical reason to emphasize the centrality
of transference interpretation. As Fairbairn (1944) pointed out, if the
therapeutic action of psychoanalysis consists of interpretation of drives
and defenses, as is claimed by the drive-ego model, it should not be
crucial whether wishes are made conscious within or outside the rela
tionship with the analyst. The acceptance of the technical principle that
the patient's deficits and conflicts are best addressed and resolved
within the analytic relationship is a tacit admission that they are
imbedded in object relationships. From the object relations perspective,
the patient's psychological structure, being a product of early attach

ments, is expected to manifest itself most clearly in its current relation
ships, most notably in the interaction with the analyst. Since pathology
is viewed as a product of relationships as they are internalized and
structured within the patient's personality, the clinical focus on the
transference is theoretically grounded in the object relations perspective
on development and pathology.
Although Sullivan (1953) and Levenson (1991) view pathology as a
matter solely of external, interpersonal relationships and eschew the
concepts of internalization and psychological structure, other aspects
of interpersonal theory are compatible with the object relations view
point. Mitchell's (1988) relational/conflict model views the mind as a
social product derived from transactional patterns and embodied in
an internalized structure that patients bring into the analytic setting.
Greenberg (1991), viewing psychological structure as a product of
the dual drives of effectance and safety, believes that the patient's
experience of safety with the analyst allows access to expanded self
representations. Gill's (1981) and Hoffman's (1991) concept of the
transference also appears to be rooted in a view of man as bringing
patterns of relating into the analytic relationship.
The object relations concept of transference is best appreciated by
contrasting it to the transference as presented by the drive-ego model.
Freud's (1912, 1915d) view was that the frustrated portion of drives
seeks new objects which in later life become the targets upon which
templates of the earlier frustrated objects are placed. In the Dora case,
Freud (1905b) recognized that the patient notices some detail in the
analyst's behavior that is similar to that of the past frustrating figure
and that becomes the stimulus for the transference perception. None
theless, Freud did not develop this idea into a broadened concept of
the transference but viewed the analyst's behavior as only the stimu
lus for the projection of a preexisting template of an early object of
frustration onto the figure of the analyst. Because, in Freud's view,
drive frustration is based largely in the Oedipus complex, the transfer
ence tends to take on the character of the frustrations of the unre
solved portions of the oedipal struggle. Thus, he conceptualized
Dora's transference as solely her frustrated erotic feelings for her
father projected onto Freud. Later, as Freud (1920) modified his drive
theory to include aggression, the concept of transference was
expanded to involve both libidinal and aggressive frustrations, but the
idea that transference is rooted in the frustration of endogenous drives
was not altered.
As we saw in chapter 1, when Anna Freud (1936) and the ego
364 Chapter 8
psychologists added the view that the patient's defenses against the
drives also become manifest in the transference, defense interpretation
became a major part of transference analysis. Despite this emendation,
the motor of the transference continues to be frustration as only those
aspects of the relationship with the analyst infiltrated by drive frustra
tion and defense against it compose the transference and are, therefore,
interpretable.
This view of the transference is narrow and reductionistic compared
to its object relations counterpart. Frustration is not a necessary compo
nent of object relationships, as object contact is motivated by the early
need to attach. The transference relationship is not reducible to frus
trated drives, their psychological expression in wishes, or any other
more primary motivation. Freed from a conflict-based theory of mental
structure, the object relations paradigm opens exploration of all early
attachments, whether conflictual or not, as contributions to self form
ation. Although differing on details, all object relations theories view
the child as navigating early relationships with the minimum degree
of anxiety possible to form a sense of self. These early bonds, with
their associated affects, form the basis of the transference. In the
object relations paradigm, the arena of transference interpretation is
broadened significantly, since every aspect of the way the patient
relates to the analyst is potentially interpretable as a manifestation of
character structure.
To choose but two examples of this broadened concept of the trans
ference, let us consider briefly Winnicott's (1963c) case of Miss X and
Kohut's (1971) treatment of Mr. A. As reported in chapter 4, Miss X's
therapeutic progress began in one crucial session during which she
lay covered with a rug and said very little. The decisive event was
Winnicott's understanding of what she needed without her having to
tell him. Miss X needed Winnicott to "take over" her omnipotence,
which allowed her the "disintegration" she needed to make con
tact with her "true self." Miss X transferred an early object relation
ship that had arrested the development of her self and could not be
reduced to a libidinal or aggressive drive.
As we saw in chapter 6, the transference of Mr. A consisted of two
primary demands: that the analyst share his values and confirm
"through a warm glow" that the patient had lived up to the analyst's
standards. When either of these needs was not met, Mr. A felt empty
and trite no matter how moral his behavior was. Kohut concluded
that the patient suffered from a "diffuse narcissistic vulnerability" due
to a gap in the idealization of the superego, which Kohut believed was
due to his thwarted efforts to idealize his father. The object relations
pattern with the father was repeated in the patient's attempts to

fill the gap in his psyche by receiving the "warm glow" from the
analyst. As in Winnicott's case of Miss X, the needs expressed as
the crucial components of the transference cannot be reduced to
endogenous drives in any meaningful sense. Both cases illustrate
the broadened, relational view of the transference that typifies the
object relations paradigm.
The much narrower drive-ego model of transference separates it
from character structure. From the object relations perspective, charac
ter structure and transference are indistinguishable. Character is
equivalent to the structure of the self, an organization of object rela
tionships that defines both the perceptions of and interactional pat
terns with others (Summers, 1993a). Transference is the character
pattern, or self structure, as it gains expression in the analytic setting.
Each object relations theory emphasizes different types of object
relations in the formation of the self and has a distinct view of how
this process occurs. Consequently, they all emphasize different factors
as crucial to the transference. For Fairbairn (1958), the transference is
the patient's effort to bring the analyst into the closed system of inner
reality formed to protect the patient against object contact. Guntrip
(1969) equated transference with the regressed needs of the patient's
ego. In Klein's (1952c) and Kernberg's (1976) views, the vicissitudes of
the aggressive drive are the crucial feature. Winnicott (1963c) saw
transference as the manifestation of early dependence needs. Kohut
(1984) and the self psychologists view transference as the manifesta
tion of early needs for grandiosity and idealization in the analytic set
ting. Each theorist emphasizes a different type of object relationship as
critical to the transference, but all conceive of transference as early
forms of object attachment manifested in the relationship with the
analyst. Because character structure is a product of object relation
ships and the attachments to which they give rise, transference is not
conceived from the object relationship perspective as projection upon
a blank screen. As Gill (1981) has pointed out, the patient's attach
ments press for interaction with the analyst, since the form of attach
ment evokes an affective reaction from the analyst, and this response
is a fundamental part of the object relationship. We have seen this
shift in the conceptualization of the transference in several object rela
tions theories. The major conceptual breakthrough in this shift to an
interactional concept of analysis was the Kleinian concept of projec
tive identification. As we saw in chapter 3, Rosenfeld (1978), Segal
(1981), and Bion (1962) all transformed Klein's intrapsychic concept
into an interactional process in which the patient evokes experiences
366 Chapter 8
in the analyst in order to communicate nonverbalized parts of

the self and get rid of its unwanted aspects along with anxiety-pro
voking affects. Racker (1968) added the distinction between com
plementary and concordant countertransference to capture the
difference between (1) the analyst as object and patient as self and
(2) the analyst as self and patient as object. Kernberg (1988) consid
ers the use of projective identification to be one of the primary clini
cal implications of the object relations approach. Winnicott (1947,
1960c) also made use of projective identification; he felt that many
patients communicate best by what they make the analyst feel and
that sometimes it is useful to tell patients what feelings they have
evoked. While Kohut did not make use of the analyst's responses to
the patient in his clinical theory, some of his followers, most notably
Stolorow, Brandchaft, and Lachmann (1988), have expanded self
psychology into an intersubjective model in which the analyst's
feelings play a prominent role in the treatment.
Kernberg (1988) has drawn out the clinical implications of the
view that children internalize object relations units in which self
and object can become reversed. Since the unit is internalized,
rather than simply the object, the patient learns to operate accord
ing to the structure of the object relationship rather than on the
basis of only one side of it. Consequently, patients may experience
themselves as either the self or object part of an object relations
unit. For example, the patient who frequently feels victimized may
wantonly disregard the analyst and attempt to defeat the treat
ment, thus reversing roles by becoming the abuser and victimizing
the analyst. Such a patient may have a conscious self-concept of a
victim but in fact has a "victim-abuser" object relations unit as a
primary component of the self structure and may enact either role.
This view of the countertransference is another way in which the
object relations paradigm broadens the concept of transference; it
alerts analysts to their affective responses to the patient and the
need to use them to understand the patient's psychological struc
ture. How the analyst feels in response to the patient is always a
significant component of the transference, and sometimes the best
clue to its nature.
The interpersonal concept of transference is a primary point of con
nection between the object relations paradigm and interpersonalist the
ory. Indeed, in this sense, the object relations approach is closer to the
interpersonalist position than the classical model. Mitchell (1988),
Greenberg (1991), Gill (1981), and Hoffman (1991) all view the analyst's
reactions to the patient as a critical component of the transference.
Although patients defend against the awareness of their object rela

tionships, it is not this type of defense that is the biggest obstacle
to therapeutic success. As every therapist knows, pointing out the
patient's object relations structure, even if accurately timed in the
transference, does not mean that the structure will change. It is com
mon for patients to see and even understand their experience of the
analyst without being able to change it. Patients tend to be reluctant to
yield any aspect of their pathology, and this resistance to change is
especially true for the structure of the self. For patients to give up their
object relations structure, they must yield their sense of self, of who
they are. It is the patient's inability to modify his or her characteristic
pattern of relating to the analyst despite insight into the pattern that
constitutes resistance from the object relations perspective (Summers,
1993a). This position contrasts with the classical point of view, accord
ing to which resistance is defense against awareness of either wishes
toward the analyst or the affect associated with them.
It should be emphasized that within the classical framework it has
always been difficult to account for the phenomenon of the patient's
refusal to change after awareness of affect. Freud (1937) attempted
to account for it primarily with his concepts of the constitutional
strength of the instincts and the deformation of the ego. He felt that
the more pathology is due to the constitutional strengths of the
instincts and the more the ego needs to be altered, the more difficult to
change process will be. Freud (1937) considered the following to be
the most recalcitrant obstacles to analytic success: (1) adhesiveness of
the libido, the inability of the patient to withdraw investment in the
object; (2) masochism, or the desire to be ill, which Freud attributed to
the death instinct; (3) the fixity of the defenses, which leads to the pur
suit of situations to justify the defenses, a phenomenon Freud called
"resistance to resistance"; (4) the rigidity inherent in all mental
processes, which Freud felt had unknown causes; and (5) the male's
repudiation of femininity and the female's penis envy. Freud was
quite pessimistic about the possibility of changing any of these phe
nomena inasmuch as the "quantitative factor" of the instincts was con
sidered unalterable and the ego not amenable to change. Indeed,
Freud considered most of these factors constitutional, and he offered
little hope for altering those that are not, such as resistance to resis
tance. With these intractable resistances, Freud seemed to feel he had
come upon psychological rock bottom.
The object relations belief in the amenability of resistance to psy
choanalytic intervention highlights an advantage of the object rela
tions paradigm as a pure psychology. From this perspective, the
368 Chapter 8
patient's attachment to the analyst is the way the self structure is

maintained; the defenses are resilient because they are protecting the
self. The patient is resistant to change even with awareness because
awareness does not change the need to maintain the sense of self. To
give up the attachment to the object is to face the feeling of not know
ing who one is, that is, of annihilation anxiety. Two critical interpre
tive implications follow. First, the fact that the resistance is an object
relationship makes it interpretable as a need to attach in a particular
manner. The object relations paradigm approaches resistance by
investigating the patient's need to form this type of relationship with
the analyst, rather than viewing it as a constitutional given. Mitchell
(1988) has pointed out the therapeutic power of patient and analyst
exploring the question, Why is this the only way of relating to the ana
lyst? Whereas the drive model hits analytic rock bottom, that is, reaches
issues deemed unanswerable through analysis, object relations theo
ries pursue analysis of the need for a given type of object relationship
and the anxiety of giving it up. In the example presented earlier of
the internalization of the victim-abuser object relationship unit, the
analytic inquiry pursued why the patient could only experience the
analyst as abuser or victim. There is no biological bedrock beyond
which the analysis cannot be pursued.
This inquiry is of course likely to run into further defenses, and this
renewed resistance leads to the second critical interpretive implication
of the object relations view of resistance. The recalcitrance of the object
relations structure, as noted, is due to the anxiety of yielding the sense
of self and bearing the consequent annihilation anxiety. The working-
through process from this perspective implies a continual confronta
tion with the threat to the sense of self during the process of structural
change. All movement toward structural change evokes anxiety of
loss of self and leads to the desire to conserve the old structure.
Working through, from the object relations perspective, is the contin
ual interpretation of this conservative tendency of the psyche. When
the patient regresses from apparent therapeutic progress, the analyst
identifies the primary cause of the regression as the anxiety of losing
the sense of self. This working-through process of confrontation with
the borders of the self continues until a new self is solidified. Thus, the
confrontation and management of the anxiety evoked by the threat to
the patient's object relations structure become crucial components of
all psychoanalytic treatment based on the object relations perspective,
irrespective of the severity of pathology.
These principles are illustrated in the analysis of the real estate
developer discussed earlier. The analysis became especially painful to
the patient when, after about one year of therapy, he became aware of
how much he had compromised himself in his desperate efforts to
maintain a relationship with his father. He had tolerated sustained,
humiliating verbal abuse in order to maintain a sense of connection
with his father and in the vain hope of winning his approval. Working
for his father, the patient had felt subjected to the whims of an almost
totalitarian presence from which he could not break loose. As he
began to become aware of the intensity of his former need for his
father, the patient acknowledged that he had begun "obsessing" about
the relationship with the analyst, which he found abusive and humili
ating. He feared that he was becoming excessively dependent on the
analyst and that he had made himself vulnerable only to meet with no
support. The patient felt intense anger at the analyst and concluded
that it would be best to leave the relationship to arrest that process.
The patient agreed with the interpretation that the analysis was being
experienced as a repetition of his relationship with the father but said
that it was for that very reason that the process was debilitating and
that he must leave it. He then acknowledged that he had drunk exces
sively when he worked for his father to make the situation "tolerable"
and maintained that the difference between analysis and the relation
ship with his father was that he no longer drank heavily; conse
quently, the analytic relationship was intolerable. He believed that
in the analysis he was "throwing [himself] at a wall," just as he had
with his father, to no avail. Remaining in analysis despite the lack of
support from the analyst was experienced as futilely groveling to
please his unresponsive father, and he did not wish to be humiliated
a second time. While fully acknowledging that he was escaping the
analytic relationship, the patient was convinced that flight from the
relationship was his only self-respecting option.
The analytic process reached a critical juncture at this point. The
patient ultimately decided to stay once he realized that fleeing was
not an escape from his father but continued imprisonment in his emo
tional grip, which reaffirmed his pattern of relating only as a humili
ated victim. Further, the patient realized that he had constructed the
analytic situation as a stricture of unrelenting hard work according to
which relaxation was "against the rules." Analysis had become
another excessive life burden. His desire to flee from this internal
prison led to an understanding of his life pattern of constant traveling.
He now realized that many of his frequent trips were not business
necessities but "escapes." His experience of himself as a victim at the
hands of his business associates and now his analyst was clearly
rooted in the paternal transference. It became evident at this point that
370 Chapter 8
continual escape was a part of the pattern. The patient acknowledged

his terror of dependence on the analyst, which he associated with his
futile attempts to win his father's love. His inability to leave his father
had led to feelings of humiliation and guilt; he had compromised him
self to the point of self-loathing out of fear of losing the connection.
His escapes through traveling and alcohol were his efforts to maintain
the self he had created of a victimized coward without becoming fully
aware of its existence.
In the next session the patient reported that after the previous hour
he had had the fantasy of being accosted by muggers, seizing their
gun, shooting their legs out from under them as they ran away, and
permanently damaging them. He connected this rage to his desire for
vengeance against his father for having put him in the humiliating
position of compromising so much in his effort to achieve a relation
ship. While he acknowledged that the fantasy signified continued
anger at the analyst, he believed that this rage was about having to
enter into a new relationship to resolve his issues. The blind rage and
its injurious intent were not a product of an endogenous drive but a
product of his complex past relationship with his father and his cur
rent relationship with the analyst and could not be understood out
side the context of his willingness to pay a crippling emotional price
for contact with his father.
The therapeutic action of the analysis consisted of the patient's con
frontation with his unwillingness to yield his sense of himself as an
overburdened, victimized, undeserving failure. In acknowledging his
dependence on the analyst, he became terrified as he began to lose the
sense of who he was. He realized that his view of himself was self-
denigrating, tension producing, and self-defeating but felt that he
would rather have that than face the terror of not knowing who he
was. He feared success for its implication of a positive view of himself,
and in this fear he included a sense of betrayal of both his parents. If
he was not burdened by guilt, he experienced a terrifying loss of
direction. The therapeutic action became the patient's continual con
frontation with his fear of relating to others without being a victim or
failure; he clung to his guilt, victimization, failure, humiliation, and
self-loathing. Over and over he faced the fear of life without these
ways of relating.
In confronting the barriers of his self the patient was already form
ing a new object relationship with the analyst and thereby a new sense
of self. For the first time, he experienced a close relationship not based
on self-loathing and humiliation and from which he did not need to
escape. Many of his self-defeating, self-denigrating life patterns began
to fall away as his sense of self started to crystallize in a new direc

tion. At this point the patient began to discard many of his former
business associates and demeaning business arrangements in favor
of more lucrative, self-respecting ventures. He also began to experi
ence both the analysis and external world with increasing joy and
relaxation. In Winnicott's (1971) terms, he was more able to "play."
This process of therapeutic change, this restructuring of the self
organization so that the various symptoms and pathological features
give way, is a manifestation of fundamental change in the patient's
object relations structure.
The patient's progress is best summarized by a dream in which he
got off a train and believed he had the wrong baggage because his
suitcase was so light and he usually "travelled] heavy." He went back
on the train to find his bags but could not; he decided he would have
to "make do" and felt the worst part would be having to buy new
underwear. The dream represents the beginning of a sense of a new,
"lighter" self that had no need to carry the older, "heavy" baggage,
but the desire to find the old, familiar baggage, reflects the self in for
mation. The new self included the ability to relax and enjoy life, as
represented by the lighter baggage. Further, the "worst part" of the
new life—"buying new underwear"—represents the rebuilding of the
self from the inside out.
It must be underscored that from the objective relations perspective
the patient's characteristic self structure promotes such functioning
as the patient is capable of. In cases of severe character pathology,
such as borderline conditions, such functioning may be tantamount to
the sense of feeling alive and real (Summers, 1988) whereas in neuro
sis the self structure is able to cope at a higher level. In the case dis
cussed here, the patient's compromise between his need to please the
admired father and his fear of threatening him with success led to
a mixture of success and self-defeat; the latter reduced the patient's
anxiety over threatening the internalized father to the point that it
allowed him to perform in many areas. That is, as long as the patient
did not enjoy too much success, he could feel satisfied with some
degree of achievement and live without crippling anxiety. The object
relations paradigm suggests that the adaptive value of this self struc
ture should be interpreted to the patient and that the transference
should be seen in this light as well. Treatment is not simply a matter
of confronting anxiety but also of interpreting to patients the reasons
why they need the types of attachments they have formed. In all these
ways, the object relations paradigm adopts a more optim istic
approach to resistance and character change than does Freud's theory.
372 Chapter 8
A major advantage of the object relations reformulation of the psycho

analytic process is the broadened interpretive field to which it gives
rise. For this reason, all object relations theories have expanded the
range of pathology considered analytically accessible to include char
acter, and even borderline, pathology. While some theories emphasize
a particular type of pathology, such as schizoid or borderline condi
tions—the Kleinians have even extended analytic treatment to psy
chotic states—all object relations theories share a paradigm that
considers the structure of the personality, even in severe character dis
orders, potentially accessible to psychoanalytic treatment. As we have
seen, the transference of the neurotic patient is as much a product of
the self structure as is that of the more severely disturbed patient.
Patients of both types are confronted with anxiety in the face of struc
tural personality change and resist change despite awareness of trans
ference motivations. The object relations paradigm opens a way
to interpret and resolve this fundamental dilemma. Freud (1937)
believed that treatment of the neurotic patient often ran into exasper
ating, perplexing, and even insurmountable obstacles when the male's
repudiation of femininity and the female's penis envy were reached.
From the object relations perspective, in both situations the patient is
holding on to a precious sense of self and attempting to avoid anxiety;
the analysis has not reached psychoanalytic bedrock but, rather, the
anxiety of changing the sense of self, and this anxiety is interpretable
to the patient. Thus the treatment of neurosis is conducted on the
same basis as the treatment of other forms of pathology.
Despite this broadened arena of interpretation, the object rela
tions paradigm rarely attributes therapeutic efficacy solely to mak
ing the unconscious conscious. Because object relations units are
formed from early attachments, they typically require a new rela
tionship for their modification. The analytic relationship is espe
cially well suited to this task, since bringing to light the unconscious
object relations structures, even with the attendant anxiety, creates
the possibility of forming a new type of object relationship with the
analyst. The object relations paradigm thus emphasizes the forma
tion of a new relationship between patient and analyst as a critical
component of the therapeutic action of the analytic process. We
now turn to this aspect of the process.
The Analytic Relationship
Since patients' pathological object relations units arose in response to

early inadequate attachments, new types of attachments present
patients with opportunities to internalize more benign relationships

and reorganize their psychological structure. Object relations theorists
differ in their conceptualization of this process, but most see the ana
lytic relationship, even if it is conceived on a purely interpretive basis,
as a new type of object contact that can be productively internalized to
form a more functional self.
The most obvious way the analytic relationship enters the treat
ment is in the interpretive process. Most object relations and interper
sonal theorists point out that the relationship established by the
analyst's interpretive posture is in itself new for the patient. Because
the patient suffers from unhealthy early relationships that endure in
the form of object relations units, this new relationship based on
understanding, should have therapeutic benefit apart from the content
of interpretations. Kohut (1984) and other self psychologists (for
example, Basch, 1985; have given the greatest weight to analytic
empathy in the formation of a therapeutically beneficial relationship;
for them, analytic understanding, including its successes and failures,
creates and sustains the relationship that is conducive to the creation
of new psychological structure.
Winnicott (1960a), like Kohut, saw a parallel to the mother-child
relationship in the analyst-patient interaction of understanding and
failing to understand. Within the interpretive model, their views are
remarkably similar. However, Winnicott (1954b, 1960a) believed that
adaptation to developmental need must often go beyond interpreta
tion to the provision of a relationship geared to developmental deficit
whereas Kohut limited the gratification of unmet needs to analytic
empathy. Nonetheless, both theorists emphasized the necessity of a
phase of illusion that is sustained by the analyst's attunement and is
only gradually relinquished, with the aid of the analyst's errors, in
favor of reality.
Fairbairn (1958) ultimately came to the view that the greatest value
of interpretation resides in its ability to break through the "closed sys
tem" of the patient to develop a new relationship. Guntrip (1969) saw
the benefit of interpretation in its ability to reach the regressed, hidden
part of the self. Even Klein (1952c), who adhered to the classical frame
work of offering only interpretations, saw as their ultimate value the
internalization of the good object. Moreover, contemporary Kleinians,
such as Rosenfeld and Segal, infer from this view of the analytic
process that the analytic relationship must play a role in treatment
in order to account for the internalization of the good object. Only
Kernberg among object relations theorists gives no role to the relation
ship beyond the content of interpretations. With this exception, then,
374 Chapter 8
object relations theorists tend to ascribe the therapeutic efficacy of

interpretations not solely to insight but also to their ability to serve as
the instrument of a new relationship for the patient.
We have seen an example of the way interpretation begins the for
mation of a new relationship in our discussion of the analysis of the
real estate developer referred to earlier. As the analysis focused on the
patient's fear of yielding the sense of himself as victimized, humili
ated, failing, and guilty, he became enmeshed in a dependent relation
ship with the analyst. A major component of the therapeutic action of
the analysis consisted of the patient's struggle to remain in this rela
tionship rather than flee it. The interpretations formed the bond that
allowed the patient to remain connected to the analyst and fostered
his dependence on the analysis. The decision of the patient, who was
tempted to flee and almost did, to remain occurred in response to the
interpretation of his pattern of escape from dependence and meaning
ful object contact. With that decision began his engagement in a new
relationship, one in which he was dependent but freer to be himself
without self-loathing and the need for escape. Thus, interpretations
were the medium through which this new relationship was realized,
but the experience of this relationship was in itself analytic progress.
Thus, it may be argued that interpretation is only one strategy for
effecting the new relationship that is crucial to the reorganization of
the personality. The view of interpretation as the instrument of a ther
apeutic relationship has led some object relations theorists to question
the primacy, even the value, of interpretation. As we saw in chapter 4,
this is exactly the conclusion Winnicott (1954b, 1960a) drew\ He de
emphasized interpretation in favor of the analyst's ability to make
contact with the arrested portion of the patient's self in order to
unblock the developmental process. According to Winnicott, this
unblocking can take place via interpretation but if other experiences
are more important for overcoming the developmental deficits, then
provision of those experiences should be given priority. Winnicott's
explicit belief that analysis is about the analyst's adaptation to, rather
than interpretation of, the patient's needs is the most radical revision
of the analytic process posited by object relations theories.
Guntrip (1969), who was strongly influenced by Winnicott, be
lieved that the therapeutic action of analysis resided in the meeting of
the patient's regressed needs, that while interpretation might be
required to arrive at the point of developmental fixation, the meeting
of the regressed needs might require noninterpretive intervention.
Fairbairn (1958) was more cautious in questioning the central role of
interpretation, but at the end of his life he seemed to be adopting the
position that the personal relationship between analyst and patient

is more important than the content of interpretation in breaking
through the patient's "closed system" of object relationships. Kohut
(1984), as we have seen, adhered to an interpretive approach but
saw the value of interpretation in the type of relationship it created.
Furthermore, Kohut believed the type of interpretation should be
limited for a prolonged period so that the therapeutic relationship
can be built. While the analytic behavior may be construed as interpre
tive in this phase, analysts' comments are limited to understanding,
with explanation kept to a minimum. Understanding has therapeutic
benefit owing to its ability to foster a new bond. Thus, while Kohut
tended to emphasize interpretation, he gave great weight to a signifi
cant period of relationship building during which time there is little
depth interpretation. Some of Kohut's followers, such as Bacal and
Newman (1990), have questioned more radically the value of interpre
tation as opposed to the creation of the optimal relationship, suggest
ing a version of self psychology that is closer to Winnicott's concept of
the analytic process.
While Klein (1952c) did not question the exclusive role of interpre
tation, she did see the aim of the analytic process to be the internaliza
tion of the good object. One may question, however, how Klein's strict
adherence to making conscious bad internalized objects can lead to
the internalization of the good object. This gap led some of her follow
ers to become more flexible in seeing the relationship as a critical part
of the treatment (Bion 1962; Rosenfeld, 1978; Segal, 1981). They tend
to locate this flexibility primarily in the projective identification pro
cess, which leads to a focus on how analysts absorb the experiences
patients project "into" them. They note that this interpersonal rela
tionship between patient and analyst is important for the internaliza
tion of the good object, the goal of Kleinian treatment. One can see in
this modification of Klein's theories a shift away from her exclusive
reliance on interpretation. Again, Kernberg is most traditional in his
view that the sole therapeutic benefit of analytic treatment resides in
the content of interpretation, even though he emphasizes the analyst's
feelings in the understanding of the patient.
It is clear, then, that the object relations theorists differ on the type
of relationship they deem most helpful to the patient and on how this
bond is formed. Nonetheless, the majority of object relations theorists
draw from the equation of pathology with a dysfunctional object rela
tions structure the implication that the relationship between patient
and analyst is critical to the therapeutic action of the analytic process.
Emphasis on the analytic relationship as the primary instrument of
376 Chapter 8
therapeutic action is another point of intersection between object rela

tions and interpersonal theorists. Interpersonal theory sees interpreta
tion as inherently relationship building because interpretation is an
interpersonal event. Even Greenberg (1991) who tends to view inter
pretation of representations as the primary therapeutic intervention,
emphasizes strongly that interpretation is an interaction. As we have
seen in chapter 7, Mitchell (1988) focuses on the analyst's expansion of
the patient's relational configurations via the relationship he or she
offers. Levenson (1991) takes an even more radical approach: he ques
tions the value of depth interpretation in favor of expanded awareness
of the applicability of relational patterns and believes change results
more from authentic encounter than from interpretation. Gill (1979,
1981) gives more importance to interpretation than most interper
sonalists, but he believes a large share of the power of interpretation
resides in the fact that it forms a new relationship and thus extri
cates the analytic relationship from the patient's relational patterns.
Hoffman (1991) tends to emphasize authentic analytic encounter as a
significant component of the process. Thus, we can see that the inter
personal theorists tend to place an even greater emphasis on the
patient-analyst relationship and less on interpretation than do object
relations theorists. The two viewpoints are alike in their rejection of
the classical model of technique, but the primary difference appears to
be in the kind of relationship deemed most conducive to therapeutic
change. The interpersonal viewpoint tends to stress authentic modes
of engagement whereas the tendency of object relations theories is to
see therapeutic change as the unblocking of developmental arrests via
the formation of a relationship that in some way offers the patient the
type of contact missing in the earlier pathogenic relationships.
According to the object relations paradigm, a component of this
relationship is the affective response of the analyst to the patient. As
mentioned earlier, the analyst's response to the patient subserves the
interpretive process. Kernberg (1988) has emphasized this aspect of
the use of countertransference. In addition, some of the Kleinians
(Racker, 1968; Rosenfeld, 1978; Segal, 1981), along with Winnicott
(1947, 1960c) and his followers (Green, 1977; Little, 1981), have
pointed out that the analyst's absorption of the feelings engendered in
him or her by the patient in itself constitutes a new relationship.
Ogden (1982) has emphasized the importance of the therapist's ability
to process and interpret the affects projected into him or her by the
patient in order to reduce the anxiety associated with troublesome
affects and allow their integration by the patient. The Kleinians' revi
sion of projective identification to an interpersonal process implies
that from the object relations perspective the analyst's affective

response is not a problematic countertransference to be avoided, but a
crucial communication from the patient that needs to be understood.
We have noted that one of the major implications of the view of
mental structure as composed of object relations units is that the
patient can be enacting either pole of such a unit while evoking the
other pole in the analyst. Such a reversal means that the analyst is
"absorbing" his or her response and using it as a communication
rather than acting on it. This therapeutic response runs contrary to the
patient's expectations based on past experience and helps promote a
new, more benign and productive, relationship. For example, if the
patient's behavior evokes a feeling of anger in the analyst, this affec
tive response leads not to hostile behavior on the analyst's part but to
the utilization of the angry response for understanding the interaction.
Interpersonal theorists, such as Mitchell (1988), Gill (1979, 1983), and
Hoffman (1991), have also stressed this point, thus providing another
connection between the interpersonal and object relations paradigms.
Gill (1979) has pointed out that the fact that the analyst responds to
the patient's troublesome affects with an interpretation is contrary to
the patient's expectations and therefore inherently beneficial. Further,
Gill's (1981) social paradigm extends the interpersonal nature of the
therapeutic process beyond projective identification, since it assumes
that the patient's material is at least partly determined by the analyst's
behavior. While the interpersonalists have made a significant contri
bution to the object relations model by emphasizing the interactional
aspects of the analytic relationship, they tend to depart more sharply
from the traditional model by conceptualizing the transference as
mutual influence and interaction (for example, Mitchell, 1988; Lev
enson, 1991; Hoffman, 1991). In the most extreme versions of this
model the analytic process seems to take on the character of an exis
tential encounter. Levenson (1991) and, to a lesser degree, Hoffman
(1991) view the analyst's authenticity as a crucial component of the
therapeutic action of analysis. Although this component of interper
sonal theory is not shared by object relations theorists, the interper
sonalists in general underscore the object relations view that the
patient communicates by interaction and that this aspect of the rela
tionship may often be the heart of the analytic process. While Mitchell
(1988) also sees a role for analytic engagement, his view of the interac
tion as a product of the patient's relational patterns is closer to the
object relations paradigm. From the latter perspective, analysts offer a
new relationship based on their assessment of the developmental fixa
tion point of the object relations structure of the patient's self. Even
378 Chapter 8
Winnicott's radical approach of giving priority to adaptation over

interpretation is based on a careful understanding of the particular
developmental arrest of each individual patient. Despite these differ
ences in emphasis, the object relations and interpersonal paradigms
are alike in departing from the classical approach by deeming the for
mation of a relationship between patient and analyst integral to the
therapeutic action of psychoanalysis.
The object relations paradigm has a further kinship with the
interpersonal model in that both are pure psychologies, in opposi
tion to the biological basis of the drive-ego model, that see the rela
tionship between patient and analyst as key to the treatment. Since
object relations structures are seen as products of interpersonal rela
tions and since these structures cannot be affected without a new
relational contact, each object relations theory has a concept of some
aspect of the self that needs to be contacted for growth to occur.
Whether the unconnected part of the self is conceptualized as narcis
sistic illusion, the true self, a schizoid core, or split-off or repressed
aggressive object relations, a major component of the self structure is
distorted and out of contact with the world. Engaging this part of the
self brings it into an interpersonal relationship that can now affect its
structure. Thus, the primary analytic task is to find a way to recog
nize and communicate with the hitherto detached part of the self. As
this engagement takes place, the crippling object relations structure
is now involved in a new, growth-enhancing interaction. The inter
nalization of this new attachment results in a new, more functional
object relations structure.
Object relations theorists have tended to see the patient as unable
to achieve personal contact in some fundamental manner. Fairbairn
(1940) emphasized schizoid withdrawal in his application of the object
relations viewpoint to the reconceptualization of psychopathology.
Winnicott (1960a, 1963b) developed the notions of self arrestation in
the "pre-reality" phases of dependence and a split between the "true"
and "false" self. Even Klein (for example, 1960), while adhering to a
drive model, saw forceful interpretation of the unconscious as the
vehicle for making contact with pathological aspects of the patient's
self. Klein's followers broadened this focus to making contact with the
withdrawn parts of the self. For Kohut, the patient's archaic grandios
ity and/or idealizing needs interfere with relationships, and the ana
lyst's task is to help the patient resume development in these areas in
the interest of forming and maintaining more mature interpersonal
contacts and adopting more realistic ambitions and goals.
As noted, the object relations paradigm of psychoanalytic treatment
is based on pure psychology. The starting point of this paradigm is the

initial attachment between child and mother; its clinical theory derives
from the fact that psychological life originates in this manner. The
human psyche is conceived to be a product of early attachments, and
psychopathology is understood to consist of defects resulting from
the absence or malignant transformation of such attachments, with
resulting problems in the internalization of healthy, self-sustaining
attachments. Each object relations theory offers its own version of de
velopment and pathogenesis and a treatment approach claiming a
degree of therapeutic success, but each view, in turn, has theoretical
and clinical weaknesses. The object relations approach is still very
much in the process of developing.
The object relations paradigm changes the interpretive model and
goes beyond it. Whether interpretation or the nature of the therapeutic
relationship is emphasized in any given object relations theory, all
such theories view the clinician's task as one of connecting with the
patient in a new way, contacting an unengaged part of the self, con
fronting the resulting annihilation anxiety, and thereby forming a new
type of attachment that eventuates in more functional object relations
structures. In this way, the implications of a pure psychology are car
ried through to the clinical process. Neither biological drives nor
abstract psychological energies nor mechanisms have a place in this
model. The patient is a person-in-interaction, and the analyst's task is
to forge a connection in order to change fundamentally the nature of
who the patient is and who that person may yet become.
References
Abraham, K. (1924), A short study of the development of the libido. In: Selected Papers on
Psycho-Analysis. London: Hogarth Press, 1954, pp 418-502.
Adler, G. (1985), Borderline Psychopathology and Its Treatment. New York: Aronson.
Alexander, F. (1950), Analysis of the therapeutic factors in psychoanalytic treatment. In:
The Scope of Psychoanalysis: Selected Papers of Franz Alexander. New York: Basic Books,
1961, pp. 261-275.
Arlow, J. (1963), Conflict, regression, and symptom formation. Internat. J. Psycho-Anal.,
44:12-22.
(1969), Unconscious fantasy and disturbances of conscious experience. Psycho-
anal. Quart., 38:1-27
(1987), The dynamics of interpretation. Psychoanal. Quart., 56:68-87.
& Brenner, C. (1964), Psychoanalytic Constructs and the Structural Theory. New
York: International Universities Press.
& --------- (1990), The psychoanalytic process. Psychoanal. Quart. 59:678-693.
Bacal, H. (1985), Optimal responsiveness and the therapeutic process. In: Progress in Self
Psychology, Vol. 1, ed. A. Goldberg. New York: Guilford Press, pp. 202-226.
& Newman, K. (1990), Theories of Object Relations: Bridges to Self Psychology. New
York: Columbia University Press.
Bakan, D. (1966), The Duality of Human Existence. Chicago: Rand McNally.
Balint, M. (1968), The Basic Fault. London: Tavistock.
Basch, M. (1984), Selfobject theory of motivation and the history of psychoanalysis. In:
Kohut's Legacy: Contributions to Self Psychology, ed. P. Stepansky & A. Goldberg.
Hillsdale, NJ: The Analytic Press, pp. 3-20.
(1985), Interpretation: Toward a developmental model. In: Progress in Self
Beebe, B. & Lachmann, F. (1992), A dyadic systems view of comm unication. In:
Relational Perspectives in Psychoanalysis, ed. N. Skolnick & S. Warshaw. Hillsdale, NJ:
The Analytic Press, pp. 61-82.
Jaffe, J. & Lachmann, F. (1992), The contribution of mother-infant influence to
the origins of self- and ob ject-rep resen tation s. In: R elational Perspectives in
Psychoanalysis, ed. N. Skolnick & S. Warshaw. Hillsdale, NJ: The Analytic Press, pp.
83-118.
Bettelheim, B. (1972), Regression as progress. In: Tactics and Techniques in Psychoanalytic
Therapy, ed. P. Giovacchini. New York: Science House, pp. 189-199.
Bibring, E. (1947), The so-called English school of psychoanalysis. Psychoanal. Quart.,
16:69-93.
(1954), Psychoanalysis and the dynamic psychotherapies. J. Amer. Psychoanal.
Assn., 2:745-770.
381
382 References
Bion, W. (1957), Differentiation of psychotic and non-psychotic personalities. In: Melanie

Klein Today: Vol. 2. Mainly Theory, ed. E. Spillius. London: Routledge, 1988, pp. 61-80.
(1959a), Attacks on linking. In: Melanie Klein Today: Vol. 1. Mainly Theory, ed. E.
Spillius. London: Routledge, 1988, pp. 87-101.
(1959b), Experiences in Groups. New York: Basic Books.
(1961), A theory of thinking. In: Melanie Klein Today: Vol. 1. Mainly Theory, ed. E.
Spillius. London: Routledge, 1988, pp. 178-186.
(1962), Learning from Experience. New York: Basic Books.
Bowlby, J. (1969), Attachment and Loss: Vol. 1. Attachment. New York: Basic Books.
Brandchaft, B. & Stolorow, R. (1984), A current perspective on difficult patients. In:
Kohut's Legacy: Contributions to Self Psychology , ed. P. Stepansky & A. Goldberg.
Hillsdale, NJ: The Analytic Press, pp. 93-116.
Brenner, C. (1976), Psychoanalytic Technique and Psychic Conflict. New York: International
Universities Press.
(1979), W orking alliance, therapeutic alliance, and transference. J. Amer.
Psychoanal. Assn., 27(suppl.):137-158.
(1981), Defense and defense mechanisms. Psychoanal. Quart., 50:557-569.
Carson, R. (1969), Interaction Concepts of Personality. Chicago: Aldine.
Celani, D. (1993), The Treatment of the Borderline Patient: Applying Fairbairn's Object
Relations Theory in the Clinical Setting. Madison, CT: International Universities Press.
Curtis, H. (1985), Clinical perspectives on self psychology. Psychoanal. Quart.,
54:339-378.
(1986), Clinical consequences of the theory of self psychology. In: Progress in Self
Fairbairn, R. (1940), Schizoid factors in the personality. In: Psychoanalytic Studies of the
Personality. London: Tavistock, 1952, pp. 3-28.
(1941), A revised psychopathology of the psychoses and psychoneuroses. In:
Psychoanalytic Studies of the Personality. London: Tavistock, 1952, pp. 28-58.
(1943), The repression and the return of bad objects (with special reference to the
"war neuroses"). In: Psychoanalytic Studies of the Personality. London: Tavistock, 1952,
pp. 59-81.
(1944), Endopsychic structure considered in terms of object-relationships. In:
Psychoanalytic Studies of the Personality. London: Tavistock, 1952, pp. 82-136.
(1946), Object-relationships and dynamic structure. In: Psychoanalytic Studies of
the Personality. London: Tavistock, 1952, pp. 137-151.
(1949), Steps in the development of an object-relations theory of the personality.
In: Psychoanalytic Studies of the Personality. London: Tavistock, 1952, pp. 152-161.
(1951), A synopsis of the development of the author's views regarding the struc
ture of the personality. In: Psychoanalytic Studies of the Personality. London: Tavistock,
1952, pp. 162-182.
(1954), Observations on the nature of hysterical states. Brit. J. Med. Psychol.,
27:116-125.
(1958), On the nature and aims of psycho-analytical treatment. Internat. J. Psycho-
Anal., 39:374-385.
Freud, A. (1927), Four lectures on child analysis. In: The Writings of Anna Freud, Vol. 1.
New York: International Universities Press, 1974, pp. 3-69.
(1936), The Ego and the Mechanisms of Defense. New York: International
Universities Press.
Freud, S. (1895), Studies on hysteria. Standard Edition, 2. London: Hogarth Press, 1955.
(1905a), Three essays on the theory of sexuality. Standard Edition, 7:135-243.
London: Hogarth Press, 1953.
References 383
(1905b), Fragment of an analysis of a case of hysteria. Standard Edition, 7:7-122.

(1912), The dynamics of transference. Standard Edition, 12:97-108. London:
Hogarth Press, 1958.
(1914), On narcissism. Standard Edition, 14:73-102.London: Hogarth Press, 1957.
(1915a), Instincts and their vicissitudes. Standard Edition, 14:117-140. London:
(1915b), Repression. Standard Edition, 14:143-158. London: Hogarth Press, 1957.
(1915c), The unconscious. Standard Edition, 14:166-204. London: Hogarth Press,
1957.
(1915d), Observations on transference love. Standard Edition, 12:157-168. London:
(1917), Mourning and melancholia. Standard Edition, 14:243-258. London: Hogarth
Press, 1957.
(1920), Beyond the pleasure principle. Standard Edition, 18:7-64. London: Hogarth
Press, 1955.
(1923), The ego and the id. Standard Edition, 19:12-59. London: Hogarth Press,
1966.
(1926), Inhibitions, symptoms and anxiety. Standard Edition, 20:87-172. London:
Hogarth Press, 1959
(1937), Analysis terminable and interminable. Standard Edition, 23:216-253.
(1940), An outline of psycho-analysis. Standard Edition, 23:144-207. London:
Geleerd, E. (1963), Evaluation of Melanie Klein's "N arrative of a Child Analysis."
Internat. ]. Psycho-Anal, 44:493-513.
Gill, M. (1954), Psychoanalysis and exploratory psychotherapy. J. Amer. Psychoanal.
Assn., 2:771-797.
(1979), The analysis of the tran sferen ce. ]. Amer. Psychoanal. A ssn., 27
(suppl.):263-288.
(1981), Analysis of the Transference, Vol. 1. New York: International Universities
Press.
(1983), The interpersonal paradigm and the degree of the therapist's involve
ment. Contemp. Psychoanal, 19:200-237.
(1991), Discussion of Psychoanalytic Theory and Its Relation to Clinical Work (by E.
Schwaber). Meeting of Chicago Psychoanalytic Assn., October, Chicago, IL.
& Hoffman, I. (1982), The analysis of transference, Vol. II. Studies of nine
audiorecorded psychoanalytic sessions. Psychological Issues, Monogr. 54. New York:
International Universities Press.
Giovacchini, P. (1979), Treatment of Primitive Mental States. New York: Aronson.
Gitelson, M. (1962), The curative factors in psychoanalysis. Internat. J. Psycho-Anal,
43:194-205.
Glover, E. (1937), Theory of the therapeutic results of psychoanalysis: A symposium.
Internat. ]. Psycho-Anal, 18:125-132.
(1945), Examination of the Kleinian System of Child Psychology. New York: Inter
national Universities Press.
Goldberg, A ., ed. (1978), Self Psychology: A Casebook. New York: International
Universities Press.
(1990), The Prisonhouse of Psychoanalysis. Hillsdale, NJ: The Analytic Press.
Gray, P. (1973), Technique and the ego's capacity for viewing intrapsychic activity. /.
Amer. Psychoanal. Assn., 21:474-494.
384 References
(1982), Developmental lag in the evolution of technique for the psychoanalysis

of neurotic conflict. J. Amer. Psychoanal. Assn., 30:621-656.
(1987), On the technique of analysis of the superego: An introduction. Psycho
anal. Quart., 56:130-154.
(1990), The nature of therapeutic action in psychoanalysis. J. Amer. Psychoanal.
Assn., 38:1083-1097.
Green, A. (1975), The analyst, symbolization and absence in the analytic setting. Internat.
J. Psycho-Anal, 56:1-19.
(1977), The borderline concept. A conceptual framework for the understand
ing of borderline patients: Suggested hypotheses. In: Borderline Personality Dis
orders, ed. P. Hartcollis. New York: International Universities Press, pp. 15-46.
(1978), Potential space in psychoanalysis. In: Between Reality and Fantasy, ed. S.
Grolnick & L. Barkin. New York: Aronson, pp. 167-190.
Greenberg, J. (1991), Oedipus and Beyond. Cambridge, MA: Harvard University Press.
& Mitchell, S. (1983), Object Relations in Psychoanalytic Theory. Cambridge, MA:
Harvard University Press.
Grolnick, S. (1990), The Work and Play of Winnicott. North vale, NJ: Aronson.
& Barkin, L. (1970), Between Reality and Fantasy. New York: Aronson.
Grotstein, J. (1986), Splitting and Projective Identification. Northvale, NJ: Aronson.
Guntrip, H. (1961a), Personality Structure and Human Interaction. New York: International
Universities Press.
(1961b), The schizoid problem, regression and the struggle to preserve an ego.
In: Schizoid Phenomena, Object Relations, and the Self. New York: International
Universities Press, 1969, pp. 49-86.
(1962), The clinical-diagnostic framework: The manic-depressive problem in the
light of the schizoid process. In: Schizoid Phenomena, Object Relations, and the Self. New
York: International Universities Press, 1969, pp. 130-164.
(1969), Schizoid Phenomena, Object Relations, and the Self. New York: International
Universities Press.
(1971), Psychoanalytic Theory, Therapy, and the Self. New York: Basic Books.
(1975), My experience of analysis with Fairbairn and Winnicott. In: Essential Papers
on Object Relations, ed. P. Buckley. New York: New York University Press, pp. 447-468.
Harlow, H. (1961), The development of affectional patterns in infant monkeys. In:
Determinants of Infant Behavior, Vol. 1, ed. B. Foss. New York: Wiley, pp. 75-88.
& Zim m erm an, R. (1959), Affectional responses in the monkey. Science,
130:421-432.
H artm ann, H. (1939), Ego Psychology and the Problem o f Adaptation. New York:
(1953), Contribution to the m etapsychology of schizophrenia. The Psycho
analytic Study of the Child, 8:177-198. New York: International Universities Press.
Kris, E. & Lowenstein, R. (1949), Notes on the theory of aggression. The
Psychoanalytic Study of the Child, 3/4:9-36. New York: International Universities Press.
Havens, I. (1976), Participant Observation. New York: Aronson.
Hendrick, I. (1942), Instinct and the ego during infancy. Psychoanal. Quart., 11:33-58.
(1943), Work and the pleasure principle. Psychoanal. Quart., 12:311-329.
Hoffman, I. (1990), In the eye of the beholder: A reply to Levenson. Contemp. Psycho-
anal, 26:291-304.
(1991), Discussion: Toward a social constructivist view of the psychoanalytic sit
uation. Psychoanal. Dial, 1:74-103.
Hughes, J. (1989), Reshaping the Psychoanalytic Domain. Berkeley: University of California
Press.
References 385
Jacobson, E. (1954), Transference problems in the psychoanalytic treatment of severely

depressed patients. J. Amer. Psychoanal. Assn., 2:595-606.
(1964), The Self and the Object World. New York: International Universities Press.
Kernberg, O. (1972), A critique of the Kleinian school. In: Tactics and Techniques in
Psychoanalytic Therapy, Vol. 1, ed. P. Giovachini. New York: Science House, pp.
62-93.
(1975), Borderline Conditions and Pathological Narcissism. New York: Aronson.
(1976), Object Relations Theory and Clinical Psychoanalysis. New York: Aronson.
(1980), Internal World and External Reality. New York: Aronson.
(1982), Self, ego, affects, and drives. J. Amer. Psychoanal. Assn., 30:893-918.
(1984), Severe Personality Disorders. New Haven, CT: Yale University Press.
(1988), Object relations theory in clinical practice. Psychoanal. Quart., 57:481-504.
, Bernstein, E., Coyne, L., Appelbaum, A., Horwitz, L. & Voth, H. (1972),
Psychoanalysis and psychotherapy: Final report of the Menninger Foundation
Psychotherapy Research Project. Bull. Menn. Clin., 36:1-275.
, Selzer, M., Koenigsberg, H., Carr, C. & Appelbaum, A. (1988), Psychodynamic
Psychotherapy of Borderline Patients. New York: Basic Books.
Khan, M. (1960), Clinical aspects of the schizoid personality: Affects and technique. In:
The Privacy of the Self. New York: International Universities Press, 1974, pp. 136-167.
(1962), Dream psychology and the evolution of the psychoanalytic situation.
In: The Privacy of the Self. New York: International Universities Press, 1974, pp.
27-41.
(1963), The concept of the cumulative trauma. In: The Privacy of the Self. New
York: International Universities Press, 1974, pp. 42-58.
(1964), Ego-distortion, cumulative trauma and the role of reconstruction in the
analytic situation. In: The Privacy of the Self. New York: International Universities
Press, 1974, pp. 59-68.
(1971), Infantile neurosis as a false self organization. In: The Privacy of the Self.
(1974), The Privacy of the Self. New York: International Universities Press.
Klein, M. (1926), The psychological principles of early analysis. In: Love, Guilt and
Reparation, 1921-1945. New York: Free Press, 1975, pp. 128-138.
(1928), Early stages of the Oedipus conflict. In: Love, Guilt and Reparation,
1921-1945. New York: Free Press, 1975, pp. 186-198.
(1933), The early development of conscience in the child. In: Love, Guilt and
(1935), A contribution to the psychogenesis of manic-depressive states. In: Love,
Guilt and Reparation, 1921-1945. New York: Free Press, 1975, pp. 262-289.
(1937), Love, guilt and reparation. In: Love, Guilt and Reparation, 1921-1945. New
York: Free Press, 1975, pp. 306-343.
(1940), Mourning and its relation to manic-depressive states. In: Love, Guilt and
(1945), The Oedipus complex in the light of early anxieties. In: Love, Guilt and
, (1946), Notes on some schizoid mechanisms. In: Envy and Gratitude, 1946-1963.
New York: Dell, 1975, pp. 1-24.
(1948a), On the theory of anxiety and guilt. In: Envy and Gratitude, 1946-1963.
New York: Dell, 1975, pp. 25-42.
(1948b), The Psychoanalysis of Children. London: Hogarth Press.
(1950), On the criteria for the termination of psychoanalysis. In: Envy and
Gratitude, 1946-1963. New York: Dell, 1975, pp. 43-47.
386
(1952a), On observing the behavior of young infants. In: Envy and Gratitude,
1946-1963. New York: Dell, 1975, pp. 94-121.
(1952b), Some theoretical conclusions regarding the emotional life of the infant.
In: Envy and Gratitude, 1946-1963. New York: Dell, 1975, pp. 61-93.
(1952c), The origins of transference. In: Envy and Gratitude, 1946-1963. New York:
Dell, 1975, pp. 48-56.
(1952d), The mutual influences in the development of the ego and id. In: Envy
and Gratitude, 1946-1963. New York: Dell, 1975, pp. 57-60.
(1957), Envy and gratitude. In: Envy and Gratitude, 1946-1963. New York: Dell,
1975, pp. 176-235.
(1958), On the development of mental functioning. In: Envy and Gratitude,
1946-1963. New York: Dell, 1975, pp. 236-246.
(1959), Our adult world and its roots in infancy. In: Envy and Gratitude,
1946-1963. New York: Dell, 1975, pp. 247-263.
(1960), Narrative of a Child Analysis. New York: Free Press, 1975.
Kohut, H. (1959), Introspection, empathy, and psychoanalysis: An examination of the
relationship between mode of observation and theory. J. Amer. Psychoanal. Assn.,
7:459-483.
(1966), Forms and transformations of narcissism. ]. Amer. Psychoanal. Assn.,
14:243-272.
(1968), The psychoanalytic treatment of narcissistic personality disorders. The
Psychoanalytic Study of the Child, 23:86-113. New York: International Universities
Press.
(1971), The Analysis of the Self. New York: International Universities Press.
(1972), Thoughts on narcissism and narcissistic rage. The Psychoanalytic Study of
the Child, 27:360-400. New York: Quadrangle Books.
(1977), The Restoration of the Self. New York: International Universities Press.
(1979), The two analyses of Mr. Z. Internat. J. Psycho-Anal. 60:3-27.
(1982), Introspection, empathy, and the semi-circle of mental health. Internat. J.
Psycho-Anal. 63:395-407.
(1984), How Does Analysis Cure? ed. A. Goldberg & P. Stepansky. Chicago:
University of Chicago Press.
& Wolf, E. (1978), The disorders of the self and their treatment: An outline.
Internat. J. Psycho-Anal, 59:413-425.
Levenson, E. (1981), Facts or fantasies: On the nature of psychoanalytic data. Contemp.
Psychoanal., 17:486-500.
(1982), Language and healing. In: Curative Factors in Dynamic Psychotherapy., ed.
S. Slipp. New York: McGraw Hill, pp. 91-103.
(1985), The interpersonal (Sullivanian) model. In: Models of the Mind: Relation to
Clinical Work., ed. A. Rothstein, New York: International Universities Press, pp. 49-67.
(1987), An interpersonal perspective. Psychoanal. Inq., 7:207-214.
(1989), Whatever happened to the cat? Contemp. Psychoanal. 25:537-553.
(1990), Reply to Hoffman. Contemp. Psychoanal, 26:299-304.
(1991), The Purloined Self. New York: William Alanson White Institute.
Lichtenberg, J. (1983), Psychoanalysis and Infant Research. Hillsdale, NJ: The Analytic
Press.
Little, M. (1981), Transference Neurosis and Transference Psychosis. New York: Aronson.
(1985), Winnicott working in areas where psychotic anxieties predominate. Free
Associations. 3:9-42.
London, N. (1985), Appraisal of self psychology. Internat. J. Psycho-Anal, 66:95-107.
Lorenz, K. (1963), On Aggression. New York: Bantam Books.
References 387
Mahler, M. (1952), On child psychosis and schizophrenia: Autistic and symbiotic infan
tile p sych oses. The P sychoanalytic Study o f the Child. 7 :2 8 6 -3 0 5 . N ew York:
(1968), On Human Symbiosis and the Vicissitudes of Individuation. New York:
(1971), A study of the separation-individuation process. The Psychoanalytic Study
of the Child, 26:403-424. New York: Quandrangle Books.
(1972), Rapprochement subphase of the separation-individuation process.
Psychoanal. Quart., 41:487-506.
(1975), On the current status of the infantile neurosis. ]. Amer. Psychoanal. Assn.,
23:327-333.
, Pine, F. & Bergman, A. (1975), The Psychological Birth of the Human Infant. New
York: Basic Books.
Masterson, J. (1976), Psychotherapy of the Borderline Patient. New York: Brunner/Mazel.
Mitchell, S. (1988), Relational Concepts in Psychoanalysis. Cambridge, MA: Harvard
University Press.
(1991), Contemporary perspectives on the self: Toward an integration. Psycho
anal. Dial., 1:121-147.
(1993), Interaction in psychoanalysis. Presented at annual spring meeting of the
Division of Psychoanalysis, American Psychological Association, New York City.
Ogden, T. (1982), Projective Identification and Psychotherapeutic Technique. New York:
Aronson.
Ornstein, P. (1978), The evolution of Heinz Kohut's psychoanalytic psychology of the
self. In: The Search for the Self, ed. P. Ornstein. New York: International Universities
Press, pp. 1-106.
& Ornstein, A. (1985), Clinical understanding and explaining: The empathic van
tage point. In: Progress in Self Psychology. Vol. 1, ed. A. Goldberg. New York: Guilford
Press, pp. 43-61.
Racker, H. (1968), Transference and Counter transference. New York: International
Universities Press.
Rapaport, D. (1951), The theory of ego autonomy. In: The Collected Papers o f David
Rapaport, ed. M. Gill. New York: Basic Books, 1967, pp. 357-367.
(1954), Clinical implications of ego psychology. In: The Collected Papers of David
Rapaport, ed. M. Gill. New York: Basic Books, 1967, pp. 586-593.
(1957), The theory of ego autonomy: A generalization. In: The Collected Papers of
David Rapaport, ed. M. Gill. New York: Basic Books, 1967, pp. 722-744.
Rayner, E. (1991), The Independent Mind in British Psychoanalysis. Northvale, NJ: Aronson.
Rosenfeld, H. (1965), Psychotic States: A Psychoanalytic Approach. London: Hogarth Press.
(1971), A clinical approach to the psychoanalytic theory of the life and death
instincts: An investigation into the aggressive aspects of narcissism. Internat. J. Psycho-
Anal, 52:169-175.
(1978), Notes on the psychopathology and psychoanalytic treatment of some
borderline patients. Internat. ]. Psycho-Anal., 59:215-221.
(1983), Primitive object relations and mechanisms. Internat. J. Psycho-Anal.,
64:261-267.
(1987), Impasse and Interpretation. London: Tavistock.
Rubovitz-Seitz, P. (1988), Kohut's method of interpretation: A critique. J. Amer. Psycho
anal. Assn., 36:933-960.
Schafer, R. (1976), A New Language for Psychoanalysis. New Haven, CT: Yale University Press.
Schwaber, E. (1979), On the "self" within the matrix of analytic theory. Internat. J.
Psycho-Anal, 60:467-479.
388 References
(1983), Psychoanalytic listening and psychic reality. Internat. Rev. Psycho-Anal.,

10:379-392.
Scott, J. (1958), Aggression. Chicago: University of Chicago Press.
Segal, H. (1980), Klein. London: Kamac Books.
(1981), The Work of Hannah Segal. New York: Aronson.
(1983), Some clinical implications of Melanie Klein's work. Internat. J. Psycho-
Anal., 64:269-276.
Spitz, R. (1965), The First Year of Life. New York: International Universities Press.
Stern, D. (1980), The early development of schemas of self, of other, and of various
experiences of "self with other." Presented at Symposium on Reflections on Self
Psychology, Boston.
(1985), The Interpersonal World of the Infant. New York: Basic Books.
Stolorow, R. (1985), Toward a pure psychology of inner conflict. In: Progress in Self
Psychology, Vol.l, ed. A. Goldberg. New York: Guilford Press, pp. 194-201.
Brandchaft, B. & Atwood, R. (1988), Psychoanalytic Treatment: An Intersubjective
Approach. Hillsdale, NJ: The Analytic Press.
& Lachmann, F. (1980), Psychoanalysis o f Developmental Arrests. New York:
Stone, L. (1954), The widening scope of indications for psychoanalysis. ]. Amer.
Psychoanal. Assn., 2:567-594
Sullivan, H. (1940), Conceptions of Modern Psychiatry. New York: Norton.
(1953), The Interpersonal Theory of Psychiatry. New York: Norton.
(1954), The Psychiatric Interview. New York: Norton.
(1956), Clinical Studies in Psychiatry. New York: Norton.
(1962), Schizophrenia as a Human Process. New York: Norton.
Summers, F. (1988), Psychoanalytic therapy of the borderline patient: Treating the
fusion-separation contradiction. Psychoanal. Psychol., 5:339-355.
(1993a), Implications of object relations theories for the psychoanalytic process.
The Annual of Psychoanalysis, 21:225-242. Hillsdale, NJ: The Analytic Press.
(1993b), Object relations theories and neurotic pathology: Toward a new para
digm. Presented at annual spring m eeting of the Division of Psychoanalysis,
American Psychological Association, New York City.
Sutherland, J. (1980), The British object relations theorists: Balint, Winnicott, Fairbairn,
Guntrip. J. Amer. Psychoanal. Assn., 28:829-860.
Tinbergen, N. (1951), The Study of Instincts. New York: Oxford University Press.
Tolpin, M. (1983), Corrective emotional experience: A self-psychological reevaluation.
In: The Future of Psychoanalysis, ed. A. Goldberg. New York: International Universities
Press, pp. 255-271.
Tolpin, P. (1988), Optimal affective engagement: The analyst's role in therapy. In:
Learning from Kohut: Progress in Self Psychology, Vol. 4, ed. A. Goldberg. Hillsdale, NJ:
The Analytic Press, pp. 160-168.
Waelder, R. (1937), The problem of the genesis of psychical conflict in earliest infancy.
Internat. J. Psycho-Anal., 18:406-473.
Weiss, J. (1971), The emergence of new themes: A contribution to the psychoanalytic
theory of therapy. Internat. J. Psycho-Anal., 52:459-467.
(1988), Testing hypotheses about unconscious mental functioning. Internat. J.
Psycho-Anal., 69:87-94.
White, R. (1963), Ego and reality in psychoanalytic theory. Psychological Issues, Monogr.
11, Vol. 3, No. 3. New York: International Universities Press.
Winnicott, D. (1945), Primitive emotional development. In: Through Paediatrics to
Psychoanalysis. New York: Basic Books, 1975, pp. 145-156.
References 389
(1947), Hate in the countertransference. In: Through Paediatrics to Psychoanalysis.

New York: Basic Books, 1975, pp. 194-203.
(1949), Mind and its relation to the psyche-soma. In: Through Paediatrics to
Psychoanalysis, New York: Basic Books, 1975, pp. 174-193.
(1950), Aggression in relation to emotional development. In: Through Paediatrics
to Psychoanalysis, New York: Basic Books, 1975, pp. 204-218.
(1951), Transitional objects and transitional phenomena. In: Through Paediatrics to
Psychoanalysis, New York: Basic Books, 1975, pp. 229-242.
(1952), Psychoses and child care. In: Through Paediatrics to Psychoanalysis, New
York: Basic Books, 1975, pp. 219-228.
(1954a), The depressive position in normal emotional development. In: Through
Paediatrics to Psychoanalysis. New York: Basic Books, 1975, pp. 262-277.
(1954b), Withdrawal and regression. In: Through Paediatrics to Psychoanalysis,
New York: Basic Books, 1975, pp. 255-261.
(1954c), Metapsychological and clinical aspects of regression within the psycho
analytical set-up. In: Through Paediatrics to Psychoanalysis. New York: Basic Books,
1975, pp. 278-294.
(1 9 5 6 a ), P rim ary m atern al p reo ccu p atio n . In: Through Paediatrics to
Psychoanalysis. New York: Basic Books, 1975, pp. 300-305.
(1956b), The anti-social tendency. In: Through Paediatrics to Psychoanalysis, New
York: Basic Books, 1975, pp. 306-315.
(1958), The capacity to be alone. In: The Maturational Processes and the Facilitating
Environment. New York: International Universities Press, 1965, pp. 29-36.
(1959), Classification: Is there a psycho-analytic contribution to psychiatric clas
sification? In: The Maturational Processes and the Facilitating Environment. New York:
International Universities Press, 1965, pp. 124-139.
(1960a), The theory of the parent-infant relationship. In: The Maturational
Processes and the Facilitating Environment. New York: International Universities Press,
1965, pp. 37-55.
(1960b), Ego distortion in terms of true and false self. In: The Maturational
1965, pp. 140-152.
(1960c), Counter-transference. In: The Maturational Processes and the Facilitating
Environment. New York: International Universities Press, 1965, pp. 158-165.
(1962a), A personal view of the Kleinian contribution. In: The Maturational
1965, pp. 171-178.
(1962b), Ego integration in child development. In: The Maturational Processes and
the Facilitating Environment. New York: International Universities Press, 1965, pp. 56-63.
(1963a), From dependence toward independence in the development of the
individual. In: The Maturational Processes and the Facilitating Environment. New York:
(1963b), The development of the capacity for concern. In: The Maturational
1965, pp. 73-82.
(1963c), Dependence in infant care, in child care, and in the psycho-analytic set
ting. In: The Maturational Processes and the Facilitating Environment. New York:
(1963d), Psychotherapy of character disorders. In: The Maturational Processes and
the Facilitating Environment. New York: International Universities Press, 1965, pp.
203-216.
References 390
(1963e), Psychiatric disorder in terms of infantile maturational processes. In: The

Maturational Processes and the Facilitating Environment. New York: International
Universities Press, 1965, pp. 230-241.
(1965), The Maturational Processes and the Facilitating Environment. New York:
(1971), Playing and Reality. London: Routledge.
Wolf, E. (1988), Treating the Self: Elements of Clinical Self Psychology. New York: Guilford
Press.
Zetzel, E. (1951), The depressive position. In: Affective Disorders, ed. P. Greenacre. New
York: International Universities Press, pp. 84-116.
(1956), An approach to the relation between concept and content in psychoana
lytic theory. The Psychoanalytic Study of the Child, 11:99-121. New York: International
Universities Press.
(1964), The analytic situation. In: Psychoanalysis in the Americas, ed. R. Litman.
Index
A 30-31, 37-38; Klein on, 88-91; Winnicott

on, 147, 152-54, 176-81
Abraham, K., 25, 28 Anality: Fairbairn on, 29-30; Klein on, 100
Abreaction, 216 Analyst's participation in the therapeutic
Adaptive point of view, 356, 371 process: Gill on, 331, 334-35, 365-66,
Addictions. See Substance abuse. 377; Hoffman on, 332, 340-41; Leven
Adler, A., 49 son on, 327, 334, 344, 376, 377;
Adler, G., 300 Stolorow on, 308-9, 335; Sullivan on,
Affective attunement in analysis, 304-5, 316
327-28 Analytic empathy: Kernberg on, 218;
Affects: Kernberg on, 192-93; Stolorow on, Kohut on, 270, 275-77, 281, 295-97,
307 301, 328, 373; Winnicott on, 163,
Agency: Bakan's concept of, 299; Klein's 169-70, 180
concept of, 89; Winnicott's concept of, Analytic "failures," Kohut on, 270, 275,
153 284, 296, 298, 300-1
Aggression: Fairbairn on, 31, 33, 40; Gun Analytic regression. See Therapeutic
trip on, 49; Hartmann on, 6; Kernberg regression.
on, 193-94, 200-2, 206, 239-43; Klein Analyzability: Fairbairn on, 43, 46; Kern
on, 73-75, 90, 120, 350, 365; Kohut on, berg's view of, 215; Klein's approach
262; neutralization of, 6; Winnicott on, to, 106; and object relations paradigm,
142-43, 145-46, 152-53, 160, 176-81; 372
Sullivan on, 314 Anger. See Aggression.
Agoraphobia: Fairbairn on, 39; Kohut on, Annihilation anxiety, 74, 347
268-69, 298 Antisocial personality: Kernberg on, 198,
Alcoholism. See Substance abuse. 201; Klein on, 92; Kohut on, 264, 265;
Alexander, F., 183 Winnicott on, 167-68, 172
Ambivalence toward objects: Fairbairn on, Anxiety of engulfment, 39
391
392 Index
Anxiety of object loss: Fairbairn on, 29, Beyond the Pleasure Principle (Freud), 73
35, 38, 39; Guntrip on, 50-52, 54, 56; Bibring, E., 106, 119
Klein on, 89, 91, 95-97, 99, 101, 111-12; Bion, W., 123-24, 127, 135, 365, 375: and
Winnicott on, 149,152 borderline psychopathology, 124; and
Anxiety as signal, 4-5, 347 development of thought, 124; and
Anxiety as threat to the self: in Klein's early mother-child interaction, 123-24;
work, 73; in Kohut's work, 254, and group process, 135; and mother as
266-68, 275, 284, 290-92, 347; in Mit container, 123-24, 127; and patient-
chell's work, 320, 342, 354, 357; in ob analyst relationship, 375; and projec
ject relations paradigm, 347, 350, 356, tive identification, 123-24, 365 , 375;
364, 368, 372; in Sullivan's work, and psychotic thought process, 124
312-15, 317, 347 Bipolar self, 252, 271, 352
Anxiety, traumatic, 347 Birth trauma, 162
Archaic infantile grandiosity: 251-52, Blank self, 189
280-81, 286, 289, 300, 350, 352. See also Blocked maturational process, 160, 353,
Omnipotence; Grandiosity; Narcissistic 374
disorders. Borderline personality organization: Bion
Arlow, J. & Brenner, C., 8, 11-12 on, 124; Green on, 188-89; Guntrip on,
"As if" personalities, 198, 201 59; Kernberg on, See Borderline per
Attachment to bad objects: Fairbairn on, sonality organization; Khan on, 185-86;
33; Guntrip on, 51, 356; Klein on, 83, Klein on, 81-83, 88; Kohut on, 254,
94 271, 299-300; Little on, 186-89; Mahler
Attachment between parent and child, 353 on, 21; and object relations paradigm,
Attachment, need for in animals and 357, 371-72; Rosenfeld on, 125, 131-32;
humans, 22, 349-51 Stolorow on, 308-9
Atwood, G., 305, 307, 358 Borderline personality organization, Kern
Authentic engagement in analysis: Hoff berg on, 197-98, 200-7, 220-30, 238-39,
man on, 332, 376; Levenson on, 241-42, 244-45: countertransference in,
327-28, 334, 376-77; Mitchell on, 321, 222-23, 229; defenses in, 202-4; denial
377 in, 201, 203, 224-25; devaluation in,
Authenticity, 166 201, 203-5, 224-25; ego weakness in,
Autistic psychosis, 20, 197 200-3, 238-39; establishment of self
Autoerotism, 249 object boundaries in, 230; fusion of
self- and object images in, 197, 200,
203, 228; integration of split object
B relations unit in the treatment of,
222-27, 229, 244-45; interpretation of
Bacal, H., 305 splitting in, 223-26, 243-45; limit setting
Bacal, H. & Newman, K., 1, 71, 247, 304, with, 224, 228-29; negative transference
373, 375 in, 220-22, 228-29, 241, 244; nonspecific
Bakan, D., 299 manifestations of ego weakness in, 201;
Balint, M., 16 observing ego in, 220, 228-29;
Barkin, L., 149 omnipotence in, 201, 203-5, 224-25; oral
Basch, M., 304, 328, 373 aggression in, 200, 202, 206, 224, 241,
Beebe, B., 348 244; paranoia in, 224; persecutory
Being-in-the world, 309-10 anxiety in, 204; positive transference
Bergman, A., 19 in, 220; primitive idealization in, 201-2,
Bettelheim, B., 184 204-5, 224; projection in, 200, 203-5,
Index 393
224; projective and introjective cycles Countertransference confession, 188, 334

in, 220, 228; projective identification, Creation of the new object: in develop
200, 203-5, 223-24; shift toward ment, 149; in psychoanalysis, 173-74,
primary-process thinking in, 201-2; 189
splitting of object images in, 200-4, Creation of the self in analysis, 174-75
206; splitting of self-images in, 200-2; Creativity, 149: in the analytic process,
structural characteristics of, 201-2; 173-75
superego pathology in, 198, 205-6, 226; Cultural experience, 149
therapeutic alliance in, 220; Curtis, H., 293, 297
transference in, 220-30
Bowlby, J., 22, 51, 70, 349, 351
Brandchaft, B., 305, 307-8, 358, 366 D
Brenner, C., 8, 11-12, 182
Death instinct: Fairbairn on, 33; Freud's
concept of, 33, 34, 351, 367; Klein's
c view of, 74
Defense analysis, 11-15
Delinquencies. See Antisocial personality.
Capacity to be alone, 151-55, 158 Delusion, as distinguished from illusion,
Carson, R., 299 165; See also Paranoia, Psychosis;
Castration anxiety: Guntrip on, 53; Klein Psychotic transference.
on, 103, 105; Kohut on, 267, 283, Denial: of dependence, 95; Kernberg on,
290-91 201, 203, 224-25; Klein on, 80-81, 88,
Celani, D., 72 91-92
Character pathology: Kernberg on, 197-98, Dependence: 27-28, 38, 40, 42; Klein on,
243; and object relations paradigm, 96; Mitchell on, 342, 354; Winnicott
357, 372; Winnicott on, 167, 176 on, 138-155
Child analysis, 107-11 Depersonalization, 159
Childhood symbiotic psychosis, 20, 197 Depression: Fairbairn on, 38; Guntrip on,
Clarification as therapeutic tool, 199, 50-51, 56-58, 68; Kernberg on, 198;
215-16, 224-25 Klein on, 93-94, 121; Kohut on, 257,
Cohesive self, 254-55, 267, 284 260, 265, 273, 284-85, 300; Stolorow
Communion, need for, 299 on, 307; Winnicott on, 178
Compromise formation, 8 Depressive anxiety: Klein's view of, 89,
Compulsive tie to the object, 83, 94 91, 95-97, 101, 111-12, 121, 122; Winni
Confrontation as a therapeutic tool, 199, cott's view of, 149, 152
215-16, 224-25 Depressive position: Fairbairn on, 29, 38,
Constructive urge, 178 350, 378; Klein on, 88-91, 118, 136, 150
Constructivism, 340-41 Derealization, 159
Corrective emotional dialogue, 304 Desire for perfection, 92
Corrective emotional experience, 304 Devaluation: Kernberg on, 201, 203-5,
Counter-projective techniques, 317 208-9, 212, 215, 224-25; Klein on, 81,
Countertransference: Green on, 189, 376; 88
Kernberg on, 217-18, 220, 222-26, 376; Development: Fairbairn on, 27-31;
Khan on, 186; Little on, 187-88, 376; Greenberg on, 333, 337; Jacobson on,
object relations paradigm in the, 17-18, 191-92; Kernberg on, 195-97;
375-76; Rosenfeld on, 127-28, 376; Klein on, 74-91; Kohut on, 249-54;
Segal on, 129, 376 Mahler on, 19-20, 192, 195; Mitchell
394 Index
on, 333, 337, 350-51; object relations and assessment, 11; and conflict-free
paradigm, 346; Sullivan on, 311-12, ego sphere, 6; and effectance motiva
333, 337 tion, 9, 323-24, 354, 363; and ego
Developmental arrest theory, 72, 269-72, autonomy, 6-10; and ego development,
282, 284, 289, 291, 295, 318, 320, 5; and ego energy, 9-10; and ego's role
335-36, 351 in the psyche, 4; Freud's concept of,
Developmental research, 22, 346-50, 352-53 16; and neutralization, 6; and object
Diagnostic assessment: ego psychological relationships, 15-23; and
view of, 11; Kernberg's approach to, psychoanalytic technique, 11-15; and
199-200 psychopathology, 4; and synthetic
Differentiation of self and object: Fairbairn function of the ego, 7; and trans
on, 28, 31-32; Jacobson on, 17; Kem- ference, 363-64
berg on, 195, 196-97; Kohut on, 250; Ego relationships, 151, 155, 168
Sullivan on, 313, 315; Winnicott on, 147 Ego splitting: Fairbairn on, 31-32, 43;
Disintegration anxiety, 254, 266-68, 275, Guntrip on, 60-61; Kernberg on,
284, 290-92, 347 196-97, 200-2; Klein on, 79-81, 82
Don Juan syndrome, 95 Ego strength, 4
Dora, case of, 363 Ego weakness: Guntrip on, 56-57, 60, 64;
Dread, 312-13, 315 Kernberg on, 5, 200-3, 215, 238-39
Drive theory: Fairbairn on, 25-26, 33, 42, Empathic attunement in the analytic pro
44; Freud on, 3, 33, 295, 345, 351; cess. See Affective attunement in
Greenberg on, 322-25; Guntrip on, analysis.
48-49, 70; Kernberg on, 192-94, 239-43, Empathy: Bacal & Newman on, 304; Basch
345-46; Klein on, 73-74; Kohut on, on, 304; Kernberg on, 218; Kohut on,
252-53, 264, 269, 290-92; Mitchell on, 247-48, 270, 275-77, 281, 295-97, 301,
318-19 306, 328, 373; Stolorow on, 308; Win
Drive-ego model, 1, 2, 6-11, 290, 293-94, nicott on, 143-44, 154, 157, 160, 352,
305-6, 330, 333, 336, 345, 347, 356, 359, 373
362-63, 365, 378 Environment, the role of, in development
Drives, modern concept of, 192, 239, 345, and pathology: Bion's view of, 123-24,
348 127; Klein's view of, 75-76; Kohut's
Drug addiction. See Substance abuse. view of, 249, 253, 311, 352; Segal's
view of, 123; Sullivan's view of, 312;
Winnicott's view of, 138, 144-46, 152,
E 154, 156
Envy, 80-81, 84-87, 91, 100, 113-17, 120,
Early parent loss, 256 125
Eating disorders: Guntrip on, 51, 58-59; Epistemology of psychoanalysis, 330, 332,
Klein on, 96, 121; Kohut on, 264 336, 338-39, 376
Effectance, drive for, 9, 323-24, 354, 363 Exhibitionistic needs, 250, 251, 255,
Ego arrest, 160, 181, 374, 378 260-63, 281
The Ego and the Id (Freud), 3
Ego ideal, 18, 192, 250-51
Ego integration: Kernberg's view of, F
196-97, 206; Klein's view of, 88-91, 118,
136, 350 Facilitating environment, 138
Ego nuclei, 140, 150 Fairbairn, W.R.D., xi, 16, 25-48, 60-61,
Ego psychology, 3-23, 48-49, 121, 363-64: 70-72, 73, 76, 83, 94, 136, 346, 350-51,
Index 395
354, 356, 365, 373, 378: and aggression, pathology, 35-38, 71-72, 376; and
31, 33, 40; and aims of psychoanalytic schizoid position, 28-29, 35; and sec
treatment, 46-47; and ambivalent ob ondary identification, 28; and shame,
ject, 38; and ambivalent phase, 30-31, 34, 347; and therapeutic action of
37-38; and anality, 29-30; and analysis psychoanalysis, 45; and transference,
of Guntrip, 47-48; and analyst as 45-46, 47, 362, 365; and transitional
beneficent parental figure, 45-46; and phase, 30-31, 38-42; and Winnicott, 45,
antilibidinal ego, 31-32; and basic en- 72
dopsychic situation, 32; and central Feminine position, 102
ego, 31-32; and dependence-autonomy Fenichel, O., 4
conflict, 71; and developmental-arrest Fetishism: Kohut on, 264-65; Winnicott
theorist, 72; and ego as subject of ex on, 165
perience, 25-26; and exciting object, Flight to the external object, 93, 96
29-30, 31, 32, 39; and Freud, 25, 32, Freud, A., 6, 11, 16, 105, 118, 363
42; and genital sexuality, 31, 40-41, Freud, S., 3-11, 22-23, 25, 48, 56, 73, 93,
42-43; and guilt, 34, 43, 347, 358; and 107, 138, 296, 329, 345, 363, 367, 372:
Guntrip, 70-72; and impulse and adhesiveness of libido, 367; and
psychology, 25-26, 42, 44; and intellec- anxiety, 4-5, 357; Beyond the Pleasure
tualization, 37; and internal bad ob Principle, 73; and death instinct, 33, 34,
jects, 29, 47; and internal saboteur, 31, 351, 367; and defenses, rigidity of, 367;
33; and internalization, 29, 47, 76; and and depression, 56; Dora, case of, 363;
interpretation, 44-45, 362, 373-74; and drive theory of, 3, 33, 295, 345, 351;
Klein, 26, 27, 28, 35, 76, 83; and and ego deformation, 367; and ego for
Kohut, 34, 45, and libidinal ego, 31-32; mation, 5; and ego psychology, 4-5;
and libido, 26-29; and Mahler, 30; and and etiology of neurosis, 3-5, 296; and
mature dependence, 30; and neurotic id, 5; and identification, 5; and innate
pathology, 38-42, 354, 356; and object aggression, 142; and instincts, constitu
seeking, 26, 350-52; and object split tional strength of, 367; Instinkte vs.
ting, 29-31; and Oedipus complex, Trieb, 193; and internalization of object
34, 40, 47-48; and oral phase, 27-28; cathexes, 5; and male repudiation of
and paranoid state, 39; and patient- femininity, 367, 372; and masochism,
analyst relationship, 45-46, 374-75; and 367; and metapsychology, 311; and
personalization of psychoanalytic mourning, 93; and narcissism, 249,
technique, 44-46; and phobias, 38-39; 251; and obstacles to analytic success,
and pleasure principle, 26-27, 49; and 367; and Oedipus complex, 5, 363; and
pleasure seeking, 27; and pream- penis envy, 367, 372; and psycho
bivalent phase of infancy, 28-29, 37; logical organization, 5; and psycho
and primary identification, 28, 38; and pathology, 4; and repression, 3-4, 33;
rejecting techniques, 29, 38; and and resistance, 367, 372; and structural
pseudoguilt, 34; and psychical divi model, 4, 32-33; and superego, 4-5, 14,
sion, 31-32; and psychoanalytic tech 33; and topographic model, 4; and
nique, 45-48; and psychoanalytic treat transference, 107, 363; and trans
ment, 42-48, 373; and psychological ference neurosis, 329; and trauma, 3;
structure, 31-34; and psychopathology, and unconscious, 3, 73
34-42; and rejecting object, 29-30, 31, Fromm, E., 49
32, 39; and repression, 31-32, 40; and Fusion of love and aggression, 177, 182
resistance, 44, 47; and schizoid
396 Index
G subjectivity in analytic theory, 188; and

Winnicott, 188
Geleerd, E., 121 Greenberg, J., 1, 23, 49, 73, 75, 137, 243,
Genital aggression, 100 246, 311, 322-25, 333-34, 336, 347, 354,
Gill, M., 182, 215, 245-46, 328-31, 334-38, 363, 366, 376: and analytic process,
343-44, 363, 365-66, 377: and analyst's 325, 333, 366, 376; and anxiety, 347;
participation, 331, 334-35, 365-66, 377; and conflict, 323-24, 354; and develop
and classical model, 330, 333-34; and ment, 333, 337; and drives, 322-25; and
constructivism, 340-41; and counter drive/structural model, 318; and dual
transference confession, 331; and drive theory, 323-24, 354, 363; and ef-
epistemological status of psycho fectance, drive for, 323-34, 354, 363;
analysis, 330; and Freud, 329; and in and ego, 324; and Fairbairn on, 322;
terpretation, 334-35, 376; and Leven and independence, need for, 323; and
son's critique of, 338-39, on object rela internal motivation, 324; and inter
tions theories, 333-34; and perspec- pretation, 334, 336, 376; Kernberg on,
tivism, 340-41; and real behavior of 318; Kohut on, 322-23; and Mahler,
analyst, 330, 377; and resistance to the 323; and Mitchell, 322-23; nonsomatic
awareness of transference, 329, 340; drive model of, 323; and object
and resistance, concept of, 329, 340; representations, 325; on relational/
and resistance to resolution of transfer structural model, 318; on relational
ence, 330, 340; and Sullivan, 329-30; theories, 322-23; and repression,
and therapeutic action of psycho 324-25; and re-representation, 324, 333;
analysis, 331, 344, 376, 377; and thera and resistance, 325; and safety, drive for,
peutic alliance, 329; and transference, 323-24, 354, 363; and safety in the thera
329-31, 334, 363, 366; and transference peutic process, 325, 363; and self-repre
neurosis, resolution of, 331 sentation, 324-25, 363; on Sullivan, 318,
Giovacchini, P., 121, 300, 339 on undirectional theory, 323; on Win
Glover, E., 118, 140 nicott, 323
Goldberg, A., 286, 301, 309-10 Grolnick, S., 137, 149
Grandiose self. See Grandiosity; Nar Grotstein, J., 121, 134-35, 177, 239: and
cissistic disorders. countertransference, 135; and projec
Grandiosity: Guntrip on, 54; Kernberg on, tive identification, 134-35; and split
260-64, 280-89, 353, 378; Klein on (om ting, 135, 177, 239
nipotence), 80-81, 88, 91; Kohut on, Guilt: Fairbairn on, 34, 43, 347, 358; Gun
250-52, 255-56, 260-65, 271, 274, 276, trip on, 57, 67, 69, 358; Klein on, 85,
278, 280-82, 286, 355; Rosenfeld on, 89-92, 94, 116, 117, 358-59; and object
125; Segal on, 125-26 relations paradigm, 347, 358-59; Win
Gratitude, 114-17 nicott on, 152-53, 176, 178, 358-59
Gray, P., 13-14 Guntrip, Harry, 48-72, 73, 75, 83, 94, 136,
Greed, 82, 91 182, 297, 354-56, 365: and active ego,
Green, A., 188-89, 337, 346, 358, 376: and 60; and aggression, 49; and antilibid-
analytic object, 189; and analytic space, inal ego, 61; and anxiety of object loss,
189; and blank self, 189; and border 50-52, 54, 56; and attachment to bad
line personality disorders, 188-89; and objects, 51, 356; and "badness" as a
creation of the new object, 189; and defense, 56; and borderline personali
countertransference, the use of, 189, ty, 59; and breast, longing for, 53-54,
376; and self, 346; and shift in the 61; and castration anxiety, 53; and
psychoanalytic paradigm, 358; and depression, 50-51, 56-58, 68; and
Index 397
desirable deserters, 50, 59, 253; and H

desire to return to the womb, 53,
57-58, 62, 69, 70; and desires for Harlow, H., 349
merger, 52-55, 355-56; and develop Hartmann, H., 6, 11, 23, 26, 138, 201: and
mental arrest theory, 72; and drive aggression, neutralization of, 7; and
theory, 48-49, 70; and eating disorders, apparatuses of primary autonomy, 6;
51, 58-59; and ego as the center of and conflict-free ego sphere, 6; and
experience, 48; and ego loss, 63-64; schizophrenic process, 7; and secon
and ego psychology, 48-49; and ego dary autonomy, 7; and undifferen
splitting, 60-61; and ego weakness, tiated ego-id matrix, 6, 26
56-57, 60, 64; and exciting versus rejec Hate in the countertransference, 176
ting bad objects, 50; and Fairbairn, Hatred, mother's for child, 176
48-49; and Freud, 48, 56-57; and goals Havens, L., 317, 334
of analytic therapy, 62-63, 373-74; and Healthy self. See Mental health.
grandiosity, 54; and guilt, 57, 67, 69, Hedonic tones in infancy, 242-43
358; and Hartmann, 48; and "in and Heidegger, M., 309-10
out" program, 52, 59, 64, 70; and in Hendrick, I., 9
fantile dependence, 51, 67; and inter Hoffman, I., 328, 331-32, 334-41, 343-44,
nalization of bad objects, 50, 60; on in 366, 376: and analyst's view of reality,
terpersonal theory, 49; and interpreta 340; and analytic change, 344; and
tion, 66-67, 373-74; and Kernberg, 54; authentic engagement, 332, 376; and
and Kohut, 49, 54, 66, 70, 297; and existential encounter, 340-41, 377; and
love as destructive, 51-52; and love Gill, 331, 339-41; on interpersonal
gone hungry, 51-52, 55, 355-56; and theory, 332; on interpretation, 334-45;
Mahler, 71; and manic-depressive and Levenson, disagreement with,
disorder, 58, 68; and narcissistic per 338-39; on Mitchell, 331; on object rela
sonality disorder, 54; and object fan tion theories, 332; on perspectivism,
tasies as defenses, 52-53; and Oedipal 332, 338, 340; on positivism, 331, 336;
conflicts, 53, 67; and Oedipal longings, and psychoanalytic epistemology, 332,
54, 61; and oral desires, 53; and 376; and psychopathology, 343; and
patient-therapist relationship, 67-68; transference, 363, 366
and preservation of the ego, 55; and Homosexual fantasies, 273-74, 278-79, 280,
psychoanalytic therapy, 62-70; and 290
psychopathology, 50-60; and psychosis, Horney, K., 49
55; and regressed ego, 60-63, 66-67, 70, Hughes, S., 121, 137
72, 365, and resistance, 65, 69; and Hypochondria: Klein on, 97, 121; Kohut
schizoid compromise, 67, 355-56; and on, 261-62, 265, 274, 284-85; Sullivan
schizoid dynamics, 51-52, 71, 355-56; on, 315
and schizoid pathology, 50-60, 71, Hypomania, 198, 201
355-56; and schizoid withdrawal, 52-54, Hysteria: Fairbairn on, 39-42; Kernberg
57-58, 71-72; and sexuality, 49; and on, 198
shame, 56-57, 60, 67, 347, 358, 373-74;
and split-off ego, 62-63; and substance
abuse, 58-59; and Sullivan, 317; and i
therapeutic regression, 62-64, 67-70;
and therapeutic stalemate, 64, 67; and Id, 25-26, 33, 43, 48, 60, 346
transference, 66, 69, 365; and Winni Idealization: Kernberg on, 196, 201-2,
cott, 63, 72 204-5, 210, 214, 224; Klein on, 80-81,
398 Index
88; Kohut on, 249-51, 255, 256, 257, 295, 300, 375; Mitchell on, 334, 376;
259-60, 262, 264-65, 350, 353, 364, 376 and object relations paradigm, 361-72;
Idealized object: Klein on, 80-81, 92, 104; Winnicott on, 163-64, 336, 373-74
Kohut on, (idealized parental imago), Intersubjective approach to self
250-52, 255-56, 271, 273, 279 psychology, 305-9, 366. See also
Idealizing transference: Jacobson on, 21; Stolorow.
Kernberg on, 230, 235-37; Kohut on, Intersubjective contexts, 306
21, 271, 272-80, 284, 287, 292 Interview technique: Kernberg on, 199;
Ideals, 251-52, 268, 306 Sullivan on, 317
Identification: Freud's concept of, 5; Jacob Introjection: Jacobson on, 17-18; Kernberg
son's concept of, 18; Kernberg on, 194; on, 194, 200; Klein on, 76-79, 80-81, 85
Klein's view of, 101; Winnicott's view Introversion, 178
of, 147 Isolation of affect, 198
Illusions of power and stature, 314, 315
Illusory object, 149, 165
Impulse-ridden personalities, 198, 201 J
Incorporation: Fairbairn on, 37-38; Kohut
on, 278 Jacobson, E.: and analyzability, 20; and
Infantile autism, 21 attunement between mother and child,
Infantile personalities, 198, 201 17-18; and development, 17-18, 191-92;
Infant's attachment to its caretaker, 348-51 and ego formation, 17-18; and ego
Inhibitions, 177, 198 ideal formation, 18; and ego identi
Intellectualization, 37, 198 fication, 18; and idealization in
Internal bad objects: Fairbairn on, 29, 47; development, 20; and idealizing trans
Guntrip on, 50, 60; Kernberg on, 194; ference, 20; and incorporation, 17; and
Klein on, 76-78 introjection, 17; and Kernberg, in
Internalization: Fairbairn on, 29, 47, 76; fluence on, 21, 191-92; and Kohut, in
Freud on, 5; Guntrip on, 50, 60; Jacob fluence on, 21; and Mahler, 19-20; and
son on, 17-18; Kernberg on, 194; Klein merger, 17; and object relationships in
on, 76-79, 111, 114, 116, 118, 136, 371, formation of psyche, 18-19, 22; and
373; Kohut on (transmuting), 250-53, oral phase, 17; and projection, 17; and
256-60, 270, 276-79, 281, 286, 289, self-object differentiation, 17-18; and
292-93, 295-96, 300 superego development, 18; and techni
Interpersonal theory of psychoanalysis, 23, que, modification of, 20; and temporal
245-46, 345, 376-78: Gill's version of, sense, 17; and treatment of depression,
328-31; Greenberg's version of, 322-25; 20
Hoffman's version of, 331-32; Leven Jaffe, J., 348
son's version of, 325-28; Mitchell's ver
sion of, 318-22; Sullivan's version of,
311-18; summary of, 332-36 K
Interpretation: ego-psychological approach
to, 11-15; Fairbairn on, 44-45, 362, Kernberg, O., xi, 1, 54, 94, 120-21, 171,
373-74; Gill on, 334-35, 376; Greenberg 177, 179, 261, 285-86, 297, 300, 307,
on, 334, 336, 376; Guntrip on, 66-67, 343, 345-46, 353, 355-59, 366, 373,
373-74; Hoffman on, 334-35; Jacobson 375-76: and abreaction, 216; and affect,
on, 20-21; Kernberg on, 199, 215-16, 192-93; and aggressive drive, 192-94,
218, 222, 361, 375; Klein on, 105-10, 239-43; and aggressiveness, excessive,
351, 378; Kohut on, 270, 272, 277, 284, 200-2, 206, 241, 243-44; and analytic
Index 399
empathy vs. maternal empathy, 218; 194, 200; and isolation of affect, 198;
and analytic regression in neurosis, and Jacobson, Edith, influence of,
217; and antisocial personalities, 198, 191-92, 238; and Kleinian School, 191,
201; and "as if" personalities, 198, 194, 212, 235, 237-38, 240; and libidinal
201; and autistic pychosis, 197; and drive, 241; and limit setting in
borderline personality organization, psychotherapy, 224; and Mahler,
197-98, 200-7, 220-30. See also Margaret, influence of, 192, 195, 238;
Borderline personality organization, and manipulation, 216; and merger,
Kernberg on; character pathology, high 214; and metapsychology, 192-94; and
and intermediate level, 198; and narcissistic personality, organization,
character pathology, low level, 197, 198-99, 204, 207, 215, 230-39, 261. See
243; and childhood symbiotic also Narcissistic personality organiza
psychosis, 197; and clarification, 199, tion, Kernberg on; and negative
215-16, 224-25; and confrontation, 199, therapeutic reaction, 226; and neurotic
215-26, 224-25; and counter- personality organization, concept of,
transference, 217-18, 220, 222-26, 376; 215; and neurotic personality organiza
and denial, 201, 203, 224-25; and tion, treatment of, 216-220, 239, 356-58;
depression, severe, 198; and and neutralization, 201; and object in
depressive-masochistic personalities, tegration in development, 195, 239,
198; and devaluation, 201, 203-5, 208-9, 244-45, 350; and obsessive-compulsive
212, 215, 224-25; developmental theory personalities, 198; and omnipotence,
of, 195-97; and diagnostic assessment, 201, 203-5, 208-9, 212, 215, 224-25; and
199-200; and differentiation of self- and paranoia, 211, 230-31, 234, 261; and
object images, 195, 196-97; and drive passive-aggressive personalities, 198;
theory, 192-94, 239-43, 345-46; and ego and persecutory anxiety, 204; and
identity, 194, 1%; and ego integration phobias, 198; and projection, 200,
in development, 196-97, 206; and ego 203-5, 214, 224; and projective iden
weakness, 200-3, 215, 238-39; Fairbairn, tification, 200, 203-5, 214, 223-24, 353,
Guntrip, critique of, 193-94, 218, 238; 366; and psychoanalysis, unmodified,
and fusion of self- and object images, 215, 216-20, 230, 238-39, 244, 246; and
197, 200, 203, 228; and growth of ego psychological structure, 194, 236; and
structure, 194, 200, 206, 346; and psychotherapy, expressive, 215, 220,
hypomanic disorders, 198, 201; and 230, 238-39, 224, 244; and
hysterical personalities, 198; and ideal psychotherapy, types of, 215; and
self- and object-representation, 196; psychotic depression, 197; and
and idealization, primitive, 201-2, psychotic transference, 201, 203, 211,
204-5, 210, 224; and idealization, types 228, 230, 234; and Racker, 217; and
of, 214; and identification, 194; and reaction formation, 198, 202; and
impulse-ridden personalities, 198, 201; repression, 196-98; and resistance, 216,
and inadequate personalities, 198; and 237; and role reversal in psychoanalytic
infantile personalities, 198, 201; and in treatment, 216-17, 222-24, 366; and
hibitions, 198; and instinct theory, Rosenfeld, 212; and sadomasochistic
modern, 192, 239; and intellectualiza- personalities, 198; and schizophrenia,
tion, 198; and internalization of object 197; and self-concept, integration of,
relationships, 194; and interpersonal 195, 350; and self-system, 195; and
theory, 245-46; and interpretation, 199, severe character pathology, 197; and
215-16, 218, 222, 361, 375; and inter sexual deviancy, 198, 201; and split
view technique, 199; and introjection, object-images, 195-200-4, 206; and split
400 Index
self-images, 195, 200-2, 208-9, 353; and 79-81, 82; and environment, role of,
suggestion, 216; and superego forma 75-77; and envy, 80-81, 84-87, 91, 100,
tion, 196, 198, 206, 359; and superego 113-17, 120; and Ema, the case of,
severity, 196, 198, 206, 226, 359; and 85-87, 95; and Fairbairn, 76, 83-84, 94,
suppression, 198; and technical 122, 136; and feminine position, 102;
neutrality, 215-16, 218; and and flight to the external object, 93, 96;
transference, 215-16, 220, 230, 235-37, and Freud, 73, 79, 93; and genital ag
243, 365, 366; and unconscious, 194; gression, 100; and gratitude, 114-17;
and Winnicott, 202, 218, 234, 238 and greed, 82, 91; and guilt, 85, 89-92,
Khan, M. M., 71, 185-86, 190, 234: and 94, 116, 117, 358-59; and Guntrip, 83,
analytic holding environment, 186; and 94, 114, 122, 136; and hypochondria,
borderline personality disorder, 185-86; 97, 121; and idealization, 80-81, 88;
and countertransference, use of, 186; and idealized object, 80-81, 92, 104;
and cumulative trauma, 185; and pro and identification, 101; and infant
tective shield, 185; and symbolic observation, 78; and innate knowledge
discourse, 186; and transitional space, of sexual intercourse, 101; and
186 internalization of bad object, 76-78;
Klein, M., xi, 16, 27, 35, 136, 297, 345-46, and internalization of good object,
353, 355-59, 365, 373, 375, 378: 78-79; and internalization of good ob
"adultomorphism" in theory of, 119; ject in treatment, 111, 114, 116, 118,
and agency, 89; and aggressive drive, 136, 371, 373; and introjection, 76-79,
73-75, 90, 120, 350, 365; and aggressivi- 80-81, 85; and Kernberg, 94; and
ty, excessive, 85-87, 355; and anal ag Kohut, 297; and libidinal drive, 73-75;
gression, 100; and analysis of children, and love relationships, 90; and low
107-11; and analysis of psychosis, 106; self-esteem, 96; and manic defense,
and analytic situation, establishment 91-93; and manic-depressive illness, 93;
of, 109,110; and analyzability, 106; and and melancholia, 93; and mourning,
annihilation anxiety, 74, 347; and "as 93; and narcissistic character disorder,
if" personality, 82; and borderline con 81, 88; and negative Oedipal position,
ditions, 81-83, 88; and castration anxie 102; and negative transference, 105-9,
ty, 103, 105; and compulsive tie to the 113-16; and neurosis, 82, 85-87, 95, 96,
object, 83, 94; and death instinct, 74; 356, 358; and object buildup, good and
and denial, 80-81, 88, 91-92; and denial bad, 75-78, 81, 102-3, 353; and object
of dependence, 95; and dependence on integration, 88-91; and object splitting,
the object, excessive, 96; and depres 35, 79-81, 82, 121, 353; and obsessional
sion, adult, 93-94, 121; and depressive neurosis, 85-87, 92-93; and Oedipus
anxiety, 89, 91, 95-97, 99, 101, 111-12, complex, 99-105, 111; Oedipus complex
121, 122; and depressive position, for girls, 101-2; Oedipus complex for
88-91, 118, 136, 350; and desire for boys, 102-5; and omnipotence, 80-81,
perfection, 92; and devaluation, 81, 88; 88, 91; and oral aggression, 100-2; and
and Don Juan syndrome, 95; and pavor noctumus, 113; and paranoia, 85,
drives, innate, 73, 74, 345-46; and 93, 95, 104; and paranoid delusions,
earliest object relationship, 74-76; and 78-81; paranoid position, 74-88, 95,
early interpretation, 107-10; and eating 102, 120; and paranoid-schizoid posi
disorders, 96, 121; and ego integration, tion, 84; and penis envy, 101; and
88-91, 118, 136, 350; and ego integra persecutory anxiety, 76-77, 79-80,
tion, arrest of, 82; ego psychological 96-97, 100-1, 111-12, 117-18, 121, 122;
critique of, 121; and ego splitting, and persecutory object, 80, 119; and
Index 401
Peter, case of, 108, 110; and play neutralizing function, 273, 278; and
technique, 106; and positive oedipal analytic empathy, 270, 275-77, 281,
position, 102; and primal scene, 87, 295-97, 301, 328, 373; and analytic
100-2, 105, 108, 111, 119; and projec "failure," 270, 275, 284, 296, 298,
tion, 75-76, 81, 85 , 88; and projective 300-1, 373; and analytic regression,
and introjective cycles, 76-78, 80-81, 85, 301; and archaic infantile grandiosity,
100, 121; and promiscuity, 95; and 251-52, 280-81, 286, 289, 300, 350, 352;
psychoanalytic technique, 105-8, 351, and autoeroticism, 249; and bipolar
378; and psychosis, 81; and regression self, 252, 271, 352; and borderline per
to the paranoid position, 95, 101; and sonality disorder, 254, 271, 299-300;
reparation, 89-91, 99, 117; and and castration anxiety, 267, 283,
resistance, 105, 109, 113-14; and 290-91; and cohesive self, 254-55, 267,
Richard, case of, 94-95, 103-5, 108, 284; and compensatory structures,
110-11; and Ruth, case of, 112-13; and 255-56; and delinquency, 264, 265; and
sadism, 86-87; and schizoid depression, 257, 260, 265, 273, 284-85,
mechanisms, 83-84, 88; and schizoid 300; and developmental arrest, 269-72,
pathology, 84, 88; and separation anx 282, 284, 289, 291, 295, 297; and
iety, 94; and severe character disintegration anxiety, 254, 266-68, 275,
pathology, 81, 87-88, 359; and 284, 290-92, 347; and drive theory, 252,
sociopathy, 92; and superego, 77-79, 264, 269, 290-92, 294-95, 297, 305-6;
99-100; and superego, arrest in and drug addiction, 264-65, 267, 302;
development of, 82; and transference, and early parent loss, 256; and eating
105-9, 365; and Trude, case of, 108-09; disorders, 264; and empathic attune-
and verbal self-abuse, 77 ment, failure in, 258, 260, 353; and em
Kleinians, 122-36, 372-73, 375, 376: and pathy and introspection as
borderline psychopathology, 124-25, psychoanalytic method, 247-48, 295,
131-32; and countertransference, 306; and explanatory phase of analysis,
133-36, 376; and depressive position, 270, 375; and exhibitionistic needs,
123; and narcissistic pathology, 125-27, 250, 251, 255, 260-63, 281; and
132-33; and paranoid position, 123; experience-near vs. experience-distant
and patient-analyst relationship, theory, 248, 297; and explanatory
127-28, 131-32, 134-36, 375, 378; and phase of analysis, 270, 375; and Fair
projective and introjective cycles, bairn, 300; and fetishism, 264-65; and
123-27, 129, 130, 131, 133, 135-36; and grandiose self, in development, 250-52,
projective identification, 123-27, 129, 255-56; and grandiose self,
133, 135-36; and psychoanalytic tech psychopathology of, 260-64, 271, 274,
nique, 123, 127-36; and psychosis, 276, 280-89, 353, 355, 378; and Gun
124-25, 127-31, 372; and transference trip, 297, 300; and healthy self, 252-53;
psychosis, 131-32; See also Bion, Grot- and homosexual fantasies, 273-74,
stein, Ogden, Racker, Rosenfeld, 278-79, 280, 290; and horizontal split,
Segal, Kernberg. 260-61, 278-79; and hypochondria,
Kohut, H., xi, 21, 49, 54, 66, 70, 71, 95, 261-62, 265, 274, 284-85; and idealiza
128, 141, 144, 164, 328, 350-53, 355-56, tion in development, 249-51, 255, 350,
358, 364, 366, 373-74, 378: and affec 376; and idealization, failure in, 256-57,
tionate and assertive feelings, 252, 266, 260, 262, 264-65, 353, 364; and
268, 291-92; and aggression, 262; and idealization of superego, 251; and
agoraphobia, 268-69, 298; and idealization of superego, failure in,
alcoholism, 300; and analyst as a 259-60, 364; and idealized parental
402 Index
imago, 250-52, 255-56, 271, 273, 279; 295, 299, 303, 353; and self-esteem,
and idealized parential imago, dis impoverishment of, 260-61, 275; and
turbances in, 257, 258-60, 262, 271, self, fragmenting, 302, 355; and self
273-74, 277; and idealizing structure, 250-53, 352; and self
transference, 21, 271, 272-80, 284, 287, selfobject relationship, 254, 268-69, 291,
292; and ideals, 251-52, 268, 306; and 295, 299, 310; and selfobjects, 249-50,
interpretation, 270, 272, 277, 284, 295, 253-54, 263, 266, 284, 292, 295, 298-9,
300, 375; and innate potential, 295; 300; and selfobjects, analytic function
and Kernberg, 214, 227, 230, 242, 261, of, 270, 272-73, 280, 282, 294-96, 355;
285-86; and Klein, 294, 297; and lust and selfobjects, distinction between
and hostility, 253, 266, 268, 291-93; archaic and mature, 293-94; and self
and maternal empathy, 249, 253, 311, objects, failure of, 255-57, 264-65,
352; and maternal empathy, failure in, 267-69, 296, 298, 302; and selfobjects,
265; and merger transference, 280, 285; need for throughout life, 293; and
and mirroring role of parental objects, selfobjects, oedipal, 266-68; and self
250, 255, 352; and mirroring failures object transference, 271-73, 279, 285,
in, 261, 263, 265; and mirroring 290-91, 293; and shame, 261-63; and
transference in broad sense, 280; and structural neuroses, pathology of, 252,
mirroring transference in the narrow 254-55, 262, 266-69, 271; and structural
sense, 280; and narcissistic behavioral neuroses, treatment of, 290-95; and
disorders, 264-66, 267; and narcissistic superego-ego ideal formation, 249-51;
libido, 251-52; and narcissistic per and talents, 252; and tension arc, 252,
sonality disorders, 254, 255-64, 271-89, 306; and Tragic Man vs. Guilty Man,
298-99, 355; and narcissistic rage, 358; and transference, 254, 270-73, 297,
262-63; and narcissistic self, the 250; 301-2, 364-65; and transference
and neutralizing capacity of psyche, neurosis, 267; and transmuting inter
251; and neutralizing capacity, failure nalization in development, 250, 256;
in, 257, 259; and nuclear self, 249, 255, and transmuting internalization, failure
260-61, 266-67, 279, 286, 295, 358; and of, 256-60; and transmuting internaliza
object libido, 251; and obsessive- tion in analytic process, 270, 276-79,
compulsive disorder, 300; and oedipal 281, 286, 289, 292-93, 295-96, 300; and
phase, 252, 266-68, 291-92; and understanding phase of analysis,
Oedipus complex, 252, 266-69, 283, 270-72, 375; and vertical split, 260-61,
291-93, 306; and optimal frustration, 263, 278-79; and Winnicott, 164, 183,
272, 277, 281, 294-96, 304-5; and oral 254, 295, 298, 373
incorporation, 278; and perversions, Kohut, H. & Wolf, E., 302
264-65, 267, 302; and phobias, 300; and Kris, E., 6
primary narcissism, 249, 251, 254-55;
and principle of the primacy of self-
preservation, 254, 269; and psychic L
structure, gaps in, 256-57; and
psychopathology, 254-69; and Lachmann, F., 305, 348, 366
psychoses, 254-55, 271, 299; and Levenson, E., 325-28, 330, 332, 334-37,
realistic ambitions, 251-52, 256, 268, 343-44: and analyst's affective attune-
289, 355; and realistic self-esteem, 289, ment, 327-28; and analyst's participa
376; and resistance, 282, 290-92, 294, tion, 327, 334, 344, 376, 377; and
297, 301; and self, cohesive, 253; and authentic engagement, 327-28, 334,
self, defect in, 266-69, 279, 290, 293, 376-77; and classical analytic model,
Index 403
326; and distortion, patient's, 338-39; and separation-individuation process,

and epistemology, 336, 338-39; and 19, 71, 323
fantasies, 326; Gill on, 331; Hoffman Manic-depressive disorder: Guntrip on,
on, 331; and interpersonal anxiety, 58, 68; Klein on, 93
326; and interpersonal patterns, Manipulation in psychotherapy, 216
delineation of, 326, 376; and interper Masochism: Kernberg on, 198; Winnicott
sonal reality, 325-26, 338-39; and inter on, 180
pretation in psychoanalysis, 376; and Masterson, J., 171, 300
Mitchell, 325; on object relations Mastery, drive for, 9-10
theories, 326; and pathology, 363; and Maternal empathy: Kohut on, 249, 253,
perspectivism, 338; on positivism, 336; 265, 311, 352; Sullivan on, 312; Win
and psychoanalysis, concept of, 343; nicott on, 143-44, 154, 157, 352
and psychoanalysis as a semiotic Maturational process, 138, 150, 157
science, 327; and reality in appearance Melancholia, 93
vs. reality behind appearance, 325; self Menninger Research Project on
psychology, 328; on Sullivan, 325-26; Psychotherapy and Psychoanalysis, 244
therapeutic action of psychoanalysis, Mental health: Kohut on, 252-53; Mitchell
327-28, 344, 376 on, 321, 353; and object relations
Libidinal stages, 26 paradigm, 356
Libido: classical view of, 4; Fairbairn on, Merger: Guntrip, 52-55, 355-56; Kernberg
26-29; Kernberg on, 241; Klein on, on, 197, 200, 203, 228; Klein on, 83,
73-75 94, 96; Kohut on, 280, 285; Mahler, 19;
Lichtenberg, 22, 119, 242, 297, 346, Segal, 126
348-49, 352-53 Merger transference: Guntrip's view of,
Limit setting in psychotherapy, 172, 224 67-70; Kohut's view of, 280, 285
Little, M., 186-88, 190: and analyst's ex Metapsychology, 192-94, 311
pression of feelings, 188; and counter Mirroring, mother's role of: Kohut on,
transference, use of, 187-88, 376; and 250, 255, 261, 263, 265, 352; Winnicott
delusional transference, 186-87; and on, 142
physical intervention, 187; and treat Mitchell, S., 1, 23, 49, 70, 71, 72, 73, 75,
ment of severely disturbed patients, 137, 243, 245-46, 311, 318-22, 325, 328,
187 330-31, 333-34, 336-37, 341-44, 347, 355,
Loewald, H., 193 363: and aims of analysis, 321, 343;
London, N., 297 and analytic tasks, 321, 344, 376; and
Lorenz, K., 345 anxiety, 320, 342, 354, 357; and
Love as destructive, 28-29, 51-52 authentic encounter, 321, 377; and
Love relationships, 90 childhood illusions, 322; and depen
Lowenstein, R., 6 dency needs, 342, 354; and develop
Lust, 253, 266, 268, 291, 293, 314 ment, 333, 337, 350-51; and develop
mental arrest theory, 318, 320, 351; on
drive model, 318-19; on Fairbairn, 318;
M and Gill 331; on Guntrip, 318; on in
fantile needs, 320; and interpretation,
Mahler, M., xi, 19-21, 22, 30, 71, 146-47, 334, 376; on Kernberg, 321; on Kohut,
183, 200: and borderline syndrome, 21; 318, 320, 321; and Levenson, 328; and
and development, 19-20, 192, 195; and mental health, 321, 353; and monadic
neurosis, 21; and psychoanalytic tech view of the mind, 318; and narcissism,
nique, 21; and psychopathology, 20-21; treatment of, 321-22; and neurosis,
404 Index
320, 342, 354, 357; and object seeking, 260-61, 262, 278-79; hypochondria in,
319, 352; and psychoanalytic process, 261-62, 265, 274, 284-85; idealization,
321, 344, 368, 377; and failure in, 256-57, 260, 262, 264-65;
psychopathology, 320-21, 341-42, 354; idealization of superego, failure of, in,
and relational/conflict model, 319, 363; 259-60;
and relational matrix theory of mind, idealized parental imago disturbances
318-19, 363, 376-77; and self-structure, in, 257, 258-60, 262, 271, 273-74, 277;
337, 350; and sexuality, 319; on idealizing transference in, 271, 272-80,
Sullivan, 318, 322; and transference, 284, 287; interpretation in treatment of,
377; and Winnicott, 318, 320 272, 277, 284; maternal empathic
Monadic view of the mind, 318 failure in, 265; merger transference in,
Mother-infant interaction, mutual regula 280, 285; mirroring failures in, 261,
tion of, 348-49 263, 265; mirroring transference in,
Mourning, 93 280; narcissistic rage in, 262-63;
Multimodal self, 306, 358 neutralizing failure in, 257, 259;
optimal frustration in the treatment of,
272, 277, 281; oral incorporation in,
N 278; perversions as a, 264-65; psychic
structure, gaps in, 256-57; resistance
Narcissistic disorders: Fairbairn on, 36-37; in, 282; self-defect in, 279; self-esteem
Guntrip on, 54; Kernberg on. See Nar impoverishment in, 260-61, 275; self
cissistic personality organization, Kern objects in, 263, 266, 284; selfobjects,
berg on; Klein, 81,88; Kohut. See Nar analytic function of, in 272-73, 280,
cissistic disorders, Kohut on; Mitchell 282; selfobjects, failure of, in, 255-57,
on, 321-22; Rosenfeld on, 125-26, 264-65; selfobject transference in,
132-33; Segal on, 125-26, 133; Winni 271-73, 279, 285; shame in, 261-63;
cott on, 159 transmuting internalization failures in,
Narcissistic disorders, Kohut on, 254, 256-60; transmuting internalization in
255-66, 267, 271-89, 298-99, 355: analyst analytic treatment of, 276-79, 281, 286,
as a neutralizing function in treatment 289; understanding phase in analysis
of, 273, 278; analytic empathy, role in of, 271-72; vertical split in, 260-61, 263
treatment of, 275-77, 281; analytic Narcissistic personality organization, Kern
"failure," role in treatment of, 275, berg on, 206-15; 230-39; 261: antisocial
284, 298; archaic infantile grandiosity, tendencies in, 207, 210-11; counter
in, 280-81, 286, 289; castration anxiety transference to, 234; depression in,
in, 283; compensatory structures in, 235-37; devaluation of analyst, 231-34;
255-56; delinquency as a, 264-65; envy in, 212, 213-15, 230-34, 236-37,
depression in, 257, 260, 265, 273, 238, 261; grandiose self in, 206-14, 230,
284-85; developmental arrest in, 282, 234, 236, 237-38; guilt in, 235-36;
284, 289; disintegration anxiety in, 275, hunger for objects in, 230; idealization
284; drug addiction as a, 264-65; early in, 213-14, 230, 235-37; idealizing trans
parent loss in, 256; eating disorders as, ference in 230, 235-37; inability to de
264; empathic failure in development pend in, 212-13, 230-34; integration of
of, 258, 260, 353; exhibitionistic needs objects in treatment of, 235, 237; inter
of, 260-63, 281; fetishism as a, 264-65; pretation in treatment of, 230-31;
grandiose self in, 255, 260-65, 271, 274, malignant narcissism in, 207, 211, 233;
276, 278, 280-89; homosexual fantasies mourning in, 235-37; narcissistic
in, 278-79, 280; horizontal split in, defenses in opposition to, 210; need
Index 405
for affirmation in, 208; negative trans mental health, 356; and motivation,
ference in, 233-35; oral aggressiveness 345, 350; and neurosis, 357-61,
in, 212, 230-32, 234-35, 237, 261; 371-72; and patient-analyst rela
paranoia in, 211, 230-31, 234, 261; tionship, 362-66, 372-79; and psycho
persecutory anxiety, in, 230; projection analysis as a pure psychology, 367;
in, 234-35; reactions to interpretations and psychopathology, 344, 356-61; and
of, 231, 237; sadism in, 233-34; split resistance, 367-72; and self, sense of,
object relations units in, 235; treatment 346, 350, 367-68; and self-structure,
of, 230-37; types of, 207 346, 350, 364, 368, 371; and
Narcissistic rage, 262-63 transference, 362-66, 371-72, 376; and
Negative oedipal position, 102 treatment, 361-79; and working
Negative therapeutic reaction, 226 through, 368-71
Negative transference; Kernberg on, Object Relations in Psychoanalytic Theory
220-22, 228-29, 241, 244; Klein on, (Greenberg & Mitchell), 1, 50
113-16 Object seeking: Fairbairn on, 26, 350-52;
Neurosis: Fairbairn on, 38-42, 354, 356; Mitchell on, 319, 352
Kernberg on, 215-220, 239, 356-58; Object splitting: Fairbairn on, 29-31, 32;
Klein on, 82, 85-87, 95, 96, 356, 358; Kernberg on, 195, 200-4, 206; Klein on,
Kohut on, 252, 254-55, 262, 266-69, 35, 79-81, 82, 121, 353; Winnicott on,
271, 290-95; Mahler, on 21; Mitchell 142
on, 320, 342, 354, 357; and object rela Obsessive-compulsive disorder: Kernberg
tions paradigm, 357-61, 371-73 on, 198; Klein on, 85-87, 92-93; Kohut
Neutralizing capacity of the psyche: Hart on, 300
mann on, 6; Kernberg on, 201; Kohut Oedipus complex: Fairbairn on, 34, 40,
on, 251, 257, 259 47-48; Freud on, 5, 363; Guntrip on,
53, 67; Klein on, 99-105, 111; Kohut
on, 252, 266-69, 283, 291-93, 306
o Ogden, T., 77, 121, 135, 376. See also Pro
jective identification.
Object instinctual needs, 144 Omnipotence: Kernberg on, 201, 203-5,
Object integration: Kernberg on, 195, 239, 208-9, 212, 215, 224-25; Klein on, 80-81,
244-45, 350; Klein on, 88-91; Winnicott 88, 91; Rosenfeld on, 125; Winnicott,
on, 140-41, 145, 150, 152, 157, 176, 158-59, 164, 373. See also Winnicott,
179-81 Infantile omnipotence.
Object libido: Fairbairn on, 26-29; Kohut Optimal affective engagement, 304
on, 251 Optimal responsiveness in the analytic
Object relations paradigm: and adaptive process, 305
point of view, 371; and analytic regres Orality: Fairbairn on, 27-28; Guntrip on,
sion, 368; and analyzability, 372; and 53; Klein on, 100-2; Kohut on, 278
anxiety, 347, 350, 356, 364, 368, 372; Ornstein, P. & Ornstein, A., 304
and attachment to early figures, 350;
and borderline personality disorder,
357, 371-72; and character pathology, p
357, 372; and countertransference,
375-76; and development, 346; and ear Paranoia: Fairbairn on, 39; Kernberg on,
ly relationships, role of, 337; and 211, 230-31, 234, 261; Klein on, 78-81,
guilt, 358; and interpersonal theory, 85, 93, 95, 104; Sullivan on, 316
375-78; and interpretation, 361-72; and Paranoid position, 74-88, 95, 102, 120
406 Index
Paranoid-schizoid position, 84 Promiscuity, 95

Passive-aggressive personalities, 198 Protective shield, mother as a, 185
Patient-therapist relationship: Bion on, Psychoanalysis as holding environment,
375; Fairbairn on, 45-46, 374-75; Gill 163-64, 168-69, 180-81, 186
on, 331, 334-35, 365-66, 377; Greenberg Psychoanalysis as pure psychology, 306,
on, 325, 363; Hoffman on, 332, 340-41, 367
376-77; Kernberg on, 218; Klein on, Psychoanalytic process, empirical study of,
109-10; Kohut on, 270-72, 275-77, 281, 14-15. See also Menninger Research
295-97, 301, 328, 373; Levenson on, Study on Psychoanalysis and
327-28, 334, 344, 376-77; Little on, Psychotherapy.
187-88; Mitchell on, 321, 377; and ob Psychoanalytic technique: Arlow & Bren
ject relations paradigm, 362-66, 372-79; ner on, 11-12; ego-psychological
Rosenfeld on, 128-29, 361, 3/3, 375; modification of, 11-15; Fairbairn on,
Segal on, 130-31, 136, 361, 373, 375; 45-48; Freud, A. on, 11; Freud, S. on,
Stolorow on, 308-9, 335 11; Gray on, 13-14; Grotstein on, 135;
Pavor noctumus, 113 Guntrip on, 62-70; Jacobson on, 20-21;
Penis envy: Freud on, 367, 372, Klein on, Kernberg on, 215, 216-20, 230, 238-39,
101 244, 246; Klein on, 105-8, 351, 378;
Persecutory anxiety: Kernberg on, 204; Levenson on, 327, 343; Mahler on, 21;
Klein on, 76-77, 79-80, 96-97, 100-1, Mitchell on, 321, 344, 368, 377; and ob
111-12, 117-18, 121, 122 ject relations paradigm, 361-79; Ogden
Persecutory object, 80 119 on, 135; Rosenfeld on, 127-29, 131-32,
Personality types in self psychology, 302-3 133; Segal on, 129-31, 133, 361; Weiss
Perspectivism, 332, 338, 340-41 on, 14-15
Perversions, Kohut on, 264-65, 267, 302 Psychological structure: ego-psychological
Phenomenological-existential view of man, view of, 3-11; Fairbairn on, 31-34;
310 Kernberg on, 196-97, 206; Kohut on,
Phobias: Fairbairn on, 38-39; Kernberg on, 250-53, 256-57, 352; object relations
198; Kohut on, 300 paradigm view of, 346-48
Pine, F., 19 Psychopathology: Fairbairn on, 34-42;
Play, 149; in the analytic process, 175, 371 Freud on, 4; Guntrip on, 50-60; Kern
Play technique, 106 berg on, 197-215; Kohut on, 254-69;
Pleasure principle, 26-27, 49 Mitchell on, 320-21, 341-42, 354;
Positive oedipal position, 102 Stolorow on, 308; Sullivan on, 315, 363
Positivism, 331, 336 Psychosis: Bion on, 124; Guntrip on, 55;
Presymbolic representational world, 348 Klein on, 81, 106; Kohut on, 254-55,
Primal scene, 87, 100-2, 105, 108, 111, 119 271, 299; Little on, 186-87; Rosenfeld
Primary narcissism, 249, 251, 254-55 on, 124-25, 127-28; Segal on, 124-25,
Projection: Kernberg on, 200, 203-5, 214, 127-28; Winnicott on, 254-55, 271, 299
224; Klein on, 75-76, 81, 85, 88 psychotherapy, types of, .715, supportive
Projective identification: Bion on, 123-29, type, 215, 230, 244
365, 375; Grotstein on, 134-35; Kem- Psychotic depression, 197
berg on, 200, 203-5, 214; Klein on, 77, Psychotic transference: Kernberg on, 201,
81, 83, 86-87, 121; Ogden on, 135, 376; 204, 211, 228, 230, 234; Litle on, 186-87
Segal on, 129-30, 133, 365, 375
Projective and introjective cycles, 76-79,
80-81, 83, 100, 121, 123-27, 129, 130,
131, 133, 135-36
Index 407
R 127-29, 131-32, 133; borderline psycho

pathology, 125, 131-32; and counter
Racker, H„ 133-35, 217-18, 337, 366, 376: transference 127-28, 376; and envy,
and complementary countertrans- 125; and narcissistic pathology, 125-26,
fererence, 134, 218, 366; and com 132-33; and noninterpretive interven
plementary identification, 134; and tion, 128-29; and omnipotent defense,
concordant countertransference, 134, 125; and patient-analyst relationship,
217-18; and concordant identification, 128-29, 361, 373, 375; and projective
134; and countertransference, 133-35, identification, 123, 125, 126-27, 365,
365, 376; and countertransference 375; and psychosis, 124-25, 127-28; and
thought vs. position, 134; and schizophrenia, the treatment of,
transference, 133 129-30; and self-object fusion, 126; and
Rapaport, D., 7, 11, 23 superego pathology, 125; and treat
Rayner, E., 16 ment of psychosis, 127-29; and Winni
Reaction formation, 198, 202 cott, 128
Realistic ambitions, development of, Rubovitz-Seitz, P., 297
251-52, 256, 268, 289, 355
Reality, development of sense of: Kohut s
on, 251-52 , 256, 268, 289, 355, 376;
and Winnicott on, 141, 143, 145-46, Sadism: Klein on, 86-87; Winnicott on, 180
148, 150, 154, 157 Sadomasochistic personalities, 198
Regression as a defense, 95, 101 Safety: need for, 323-24, 354, 363; in the
Regression in psychoanalysis. See therapeutic process, 15, 325, 363
Therapeutic regression. Sampson, H., 325
Relational matrix theory of the mind, Satisfaction, need for, 311, 315
318-319, 363, 376-77 Schafer, R., 362
Relational model, 311, 318 Schizoid, pathology: Fairbairn on, 35-38,
Reparation: Klein on, 89-91, 99, 117; 71-72, 376; Guntrip on, 50-60, 71,
Winnicott on, 178 355-56; Klein on, 84, 88
Repression: drive-ego model, in, 347; Fair Schizoid position: Fairbairn's concept of,
bairn on, 31-32, 40; Freud on, 3-4, 33; 28-29, 35; Guntrip's view of, 51-52;
Greenberg on, 324-25; Kernberg on, Klein's view of, 84
196-98 Schizophrenia: childhood, 21; Kernberg
Re-representation, 324-25, 333 on, 197; Rosenfeld on, 129-30; Segal
Resentment, 314 on, 129-30; Sullivan on, 316
Resistance: Fairbairn on, 44, 47; Freud on, Schwaber, E., 306, 335
3-4, 33, 367, 372; Gill on, 329-30, 340; Scott, J., 120, 240
Greenberg on, 325; Guntrip on, 65, 69; Security, need for, 313, 315
Kernberg on, 216, 237; Klein on, 105, Segal, H., 16, 88, 118, 123, 125-27, 129,
109, 113-14; Kohut on, 282, 290-92, 135, 136, 373, 375-76: and countertrans
294, 297, 301; and object relations ference, 129, 376; and early mother-
paradigm, 367-72; and Stolorow on, infant interaction, 123; and narcissistic
309 pathology, 125-26, 133; and patient-
Role reversal in psychoanalytic therapy: therapist relationship, 130-31, 136, 361,
Kernberg on, 216-17, 222-24, 366; 373, 375; and projective identification,
Racker on, 134, 218 129-30, 133, 365, 375; and psycho
Rosenfeld, H., 123-29, 131-33, 361, 365, analytic technique, 129-31, 133, 361
375, 376: and analytic technique, Self, defect in: Kohut on, 266-69, 279, 290,
408 Index
293, 295, 299, 303, 353; Winnicott on, Splitting of objects. See Object splitting.
157-67 Stealing. See Antisocial personality
Self, development of: Kohut on, 249-53; Stem, D., 22, 119, 146, 183, 242-43, 346,
Mitchell on, 333, 337, 350-55; Winnicott 348-49, 352, 354
on, 140-41, 145-46 Stimulus regulation in infants, 349
Self, fragmenting: Kohut on, 302, 355; Stimulus seeking in infants, 9, 346, 348
Winnicott on, 157 Stolorow, R., 71, 305, 307-8, 335, 337, 358,
Self, sense of, 306, 346-47, 350-52, 361-62; 366: and affects, 307; and analyst as a
See also Kohut; Winnicott. coparticipant, 308-9, 335; and analytic
Self-concept, 195, 350 goals, 308; and borderline concept,
Self-esteem: Kernberg on, 195, 350; Klein 308-9; critique of Kohut's concept of
on, 96; Kohut on, 260-61, 289, 376 the self, 306; and depression, 307; and
Self-object awareness, premature, 157 empathy, 308; and guilt, 307; and in
Self-object distinction, development of: terpersonal theory, 335; and intersub
Fairbairn on, 29-30; Kernberg on, 195, jective approach, 337, 366; and
196-97; Winnicott on, 143, 154 multimodal self, 306, 358; and psycho
Selfobjects, 249-50, 253-54, 263, 266, 270, analysis as a pure psychology, 306; and
272-73, 280, 282, 293, 294-96, 298, psychopathology, 308; and resistance,
302-10; See also Kohut, and selfobjects. 309; and self, concept of, 306; and
Selfobjects, failure in, 255-57, 264-655, selfobject, concept of, 307; and self
267-69, 2%, 298, 302, 307; See also object failure, 307; and selfobject func
Kohut, selfobjects, and failure of. tions, 307; and structural conflict,
Self pathology, types of, in self 307-8; and transference, 308-9; and
psychology, 302-3 transference cure, 309
Self-preservation, principle of primacy of, Stone, L., 200
254, 269 Stranger anxiety, 147, 149-50
Self psychology: in broad sense, 21, 49, Structural model, 4, 32-33, 196-97
247, 266, 269, 271, 281, 291, 293, 295, Subject-object distinction, 309
297, 302-10; in narrow sense, 247, 271 Substance abuse: Guntrip on, 58-59;
Self-representation, 324-25, 363 Kohut on, 264-65, 267, 300, 302;
Self-system, 195, 314 Winnicott on, 168
Self-structure: Kernberg on, 195-97; Kohut Suggestion, 216
on, 250-53, 352; Mitchell on, 337, 350; Suicide, 178
and object relations paradigm, 346, Sullivan, H., 23, 49, 311-18, 325-26,
350-52, 356, 361-62, 364, 371 329-31, 333-34, 362-63: and anger, 314,
Separation anxiety, 94, 147, 149 and anxiety, 312-15, 317, 347; and
Sexual promiscuity, 168 "bad m e," 313; and conjunctive
Sexuality: Guntrip on, 49; Mitchell on, motivation, 315; and counterprojective
319, Sullivan on, 314 techniques, 317; and development,
Shame: Fairbairn on, 34, 347; Guntrip on, 311-12, 333, 337; and "disintegration,"
56-57, 60, 67, 347, 358, 373-74; Kohut 312, 314; and dissociative system, 312,
on, 261-63 316; and dread, 312-13, 315; and ex
Sociopathy. See Antisocial personality. ploitive attitudes, 315; Freud on, 311,
Spitz, R., 146-47, 183 316; Gill, 330-31; and "good m e," 313;
Split-off ego (self): Fairbairn on, 32-33; and hypochondria, 315; and illusions
Guntrip on, 62-63; Kernberg on, 195, of power, stature, 314, 315; and inter
200-2, 208-9, 353; Klein on, 79, 81-82 pretation, 317, 334; and Levenson,
Splitting of the ego. See Ego splitting. 325-26; and low self-esteem, 315; and
Index 409
lust dynamism, 314; and malevolent 376; Mitchell on, 321, 377; Sullivan on,
transformation, 314; and maternal em 317; Winnicott on, 158-62, 264, 180-82,
pathy, 312; and nosological categories, 373-74, 378
315; and “ not m e /' 313, 315; and Therapeutic alliance, 319
paranoid transformation, 316; and Therapeutic regression: Guntrip on, 62-64,
psychiatric interviewing, 317; and 67-70; Kernberg on, 217; Kohut on,
psychopathology, 315, 363; and 301; Winnicott on, 160-63, 364
reflected appraisals, 313; and resent Therapist as participant. See Analyst's
ment, 314; and satisfaction, need for, participation.
311, 315; and schizophrenic process, Thinking, development of, 124
316; and security, need for, 313, 315; Tinbergen, N., 345
and security operations, 313, 315-16; Tolpin, M., 304
and self- and object-image fusion, 313; Tolpin, P., 304
and self-system, 314; and sex, 314; and Topographic model, 3-4
substitutive processes, 315; and Transference: drive-ego model in, 363,
tenderness, need for, 312, 314; and 365; Fairbairn on, 45-46, 47, 362, 365;
therapeutic strategy, 317; and therapist Freud on, 107, 329, 363; Gill on,
as participant observer, 316; and 329-31, 334, 363, 366; Hoffman on, 363,
transference, 316-17, 362; and treat 366; Kernberg on, 215-16, 220, 343,
ment goals, 316 365, 366; Klein on, 105-9, 113-16, 365;
Summers, F., 342-43, 357, 367, 371 Kohut on, 254, 270-73, 297, 301-2,
Superego: Freud's concept of, 4-5, 14, 33; 364-65; Mitchell on, 377; object rela
Gray on the analysis of, 14; Kernberg tions paradigm, in, 362-66, 371-72, 376;
on, 196, 198, 206, 226, 359; Klein on, Stolorow on, 308-9; Sullivan on,
77-79, 99-100; Kohut on, 249-51; Segal 316-17, 362
on, 125 Transference cure, 309
Superego pathology: Kernberg on, 196, Transference neurosis, 267
198, 206, 226, 359; Klein on, 82; Transitional phase: Fairbairn on, 30-31,
Rosenfeld on, 125 38-42; Winnicott on, 148-49
Suppression, 198 Transitional phenomena and objects,
Symbiosis. See Merger. 148-49, 165, 171, 297
True self-false self distinction, 161-62,
165-67, 181, 183, 364, 378
T
Talents, role in the self, 252 u

Technical neutrality, 215-16, 218
Temporal continuity, sense of, 141, 150, Unconscious, concept of, 3-4, 73, 194
151-52 ''Unthinkable'' anxiety, 140
Tenderness, need for, 312, 314 Use of therapist as an object, 172-75
Therapeutic action of psychoanalysis: Fair
bairn on, 45; Gill on, 331, 344, 376,
377; Guntrip on, 62-70; Hoffman on, w
332, 340-41; 344, 376; Kernberg on,
216-39; Klein on, 105-11, 114, 116, 118, Waelder, R., 118-19
136, 371, 373; Kohut on, 270-72, Weiss, J., 14-15, 62, 325
276-79, 281, 286, 289, 292-93, 295-96, White, R., 9-10, 346, 348-49
300, 375; Levenson on, 327-28, 344, Winnicott, D.W., 35, 63, 71, 72, 90, 116,
410 Index
128, 136, 202, 218, 234, 238, 295, 336, transference, 176; and hatred, mother's
350-51, 364, 366, 371, 373-74, 376, 378: for child, 176; and holding environ
and ability to use objects, 146; and ab ment, 145-46; and identification, 147;
solute dependence, developmental and illusory object, 149, 165; and in
phase of, 138, 139-46, 181; and addic fantile omnipotence, arrest in, 158-59;
tions, 168; and agency, 153; and and infantile omnipotence in develop
aggression in development, 142-43, ment, 140-1, 143-44, 145, 147, 154,
145-46, 152-53; and aggression, 157-60, 165, 364; and infantile omnipo
pathological expression of, 160, 176-81; tence, treatment of, 164, 373; and
and aggression, "pre-ruth" form of, infant's "destructiveness," 146; and
142, 146, 160; and aggression, split off, integration of objects in development,
176-77; and ambivalence toward whole 140-41, 145, 150, 152, 157; and integra
objects, in development, 147, 152-54; tion of objects in therapeutic process,
and ambivalence toward whole objects, 176, 179-81; and interpretation, the role
unresolved, 176-81; and analyst as a of, 163-64, 336, 373-74; and introver
"good-enough" mother, 163, 169-70, sion, 178; and Klein, 90, 136, 137; and
180; and antisocial tendency, 172; and Klein's depressive position, 152; and
authenticity, 166; and blocked matura Kohut, 141, 164, 183, 254, 295, 373;
tional process, 160, 353, 374; and and Little Margaret, analysis of,
borderline personality disorders, 159, 162-63, 180; and "lived psychic
167, 171-73, 175-76; and capacity to be reality," 141, 143, 150, 153, 159, 168;
alone, 151-55, 158; and character and "living together," 156; and
disorders, 167, 176; and character masochism, 180; and maternal
inhibition, 177; and constructive urge, empathy, 143-44, 154, 157, 352; and
178; and countertransference, 171-72, maturational process, 138, 157, 350;
175-76, 366, 376; and creation of object and mother's mirroring role, 142; and
in development, 149; and creation of narcissistic personality disorders, 159;
self in analysis, 174-75; and creativity, and object instinctual needs, 144; and
149; and creativity in analytic process, object mother, 144, 152; and objective
173-75; and cultural experience, 149; object, 146; and "personalization,"
and delusion vs. illusion, 165; and development of, 141, 145, 150, 157;
depersonalization, 159; and depres and play in analytic process, 175, 371;
sion, 178; and depressive anxiety, 149, and play in development, 149; and
152; and derealization, 159; and ego premature self-object awareness, 157;
arrest, 160, 181, 374, 378; and ego and primary maternal preoccupation,
coverage, 155; and ego needs, 144; and 143; and privation-deprivation, distinc
ego nuclei, 140, 150; and ego related tion between, 157-167; and psycho
ness, 151; and ego relationships, 151, pathology of absolute dependence,
155, 168; and ego support, 154; and 157, 67; and psychopathology of
environmental mother, 144-46, 152, relative dependence, 167-81; and
154, 350; and facilitating environment, psychopathology of transitional phase,
138, 156; and fetishism, 165; and 168, 171-72; and psychosis, 158-60; and
Freud, 138, 142; and fusion of love and reaction formation, 177; and reality
aggression, 177, 182; and "going on principle, 143; and "realization,"
being," 157; and "good-enough" development of, 141, 145-46, 148, 150,
mother, 145, 154, 156, 167; guilt, 154, 157; and relative dependence,
152-53, 176, 178, 358-59; and Guntrip, developmental phase of, 138, 142,
182; and hate in the counter 146-55; and reliving the birth trauma,
Index 411
162; and reparation, 178; and sadism,

180; and self-object distinction, 143,
154; and self, the sense of, 147, 150,
153; and separation anxiety, 147, 149;
and sexual promiscuity, 168; and split
ting of objects, 142; and stage of con
cern, 152, 359; and stealing, 167-68;
and stranger anxiety, 149-50; and sub
jective object, 146; and suicide, 178;
and temporal continuity, sense of, 141,
150, 151-52; and therapeutic adapta
tion, 158-62, 164, 181-82, 373-74, 378;
and therapeutic empathy, 160, 373;
and therapeutic "failures," 164, 373;
and therapeutic holding, 162-63, 180;
and therapeutic limit setting, 172; and
therapeutic regression, 160-63, 364;
and transitional objects, 148-49, 165,
171, 297; and transitional phenomena,
148-49; and transitional space in
development, 149, 351; and transitional
space in analytic process, 172-75, 183;
and true self-false self pathology,
161-62, 165-67, 181, 183, 364, 378; and
"unthinkable" anxiety, 140; and use of
therapist as object, 172-75
Wolf, E., 247, 264, 280, 302-3, 304
Working models of mother, 243
Working through in psychoanalysis,
368-71

Frank Summers - Object Relations Theories and Psychopathology - A Comprehensive Text-The Analytic Press (1994)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Frank Summers - Object Relations Theories and Psychopathology - A Comprehensive Text-The Analytic Press (1994)

Uploaded by

Copyright:

Available Formats

O b je c t R e l a t io n s T h e o r ie s

and P sych o path o lo g y

Published 2014 by Psychology Press

and by Psychology Press

Psychology Press is an imprint o f the Taylor & Francis Group,

Copyright © 1994 by The Analytic Press

Trademark notice: Product or corporate names may be trademarks or registered

Library of Congress Cataloging in Publication Data

ISBN 13: 978-0-881-63155-5 (hbk)

1 - The Origins of Object Relations Theories 1

2 - The Work of W. R. D. Fairbairn and Harry Guntrip 25

3 - The Work of Melanie Klein 73

4 - The Work of D. W. Winnicott 137

5 - The Work of Otto Kernberg 191

6 - The Work of Heinz Kohut 247

7 - The Interpersonalists 311

8 - An Object Relations Paradigm for Psychoanalysis 345

I wish to express my gratitude to Dr. Paul Stepansky for the excep­

theorists to enrich the reader's appreciation of the contributions of

recent years as psychoanalytic theorists, and clinicians who have

Greenberg and Mitchell do not present each theory as an integrated

and treatment. From this conceptual clarification, we can see the

THE ORIGINS OF EGO PSYCHOLOGY

When Freud abandoned the theory of sexual trauma as the etiology of

because the mechanism of repression is unconscious. These facts led

a defense mechanism to ward off the threatening affect and restore

In Freud's formulation, the id drives the organism to seek object

relinquishment of these objects, they are taken in through identifica­

THE CLASSICAL EGO PSYCHOLOGISTS

Freud's pioneering suggestions regarding the importance of the ego

two sources of ego development: the motivation of inborn appara­

compromised to satisfy the demands of guilt and anxiety, symptoms

minimal role to object relationships in the development of the ego.

EGO PSYCHOLOGY AND THE PSYCHOANALYTIC PROCESS

The classical ego psychologists tended to emphasize theory, but they

analysis by drawing out the clinical implications of the view that

group of ego psychologists believes that there has been a "develop­

EGO PSYCHOLOGY AND OBJECT RELATIONSHIPS

In contrast, the third branch of ego psychology views ego formation as

Edith Jacobson and Margaret Mahler

In America Edith Jacobson (1964) was the first theoretician to link

of ego development extended the role of object relationships well

provided no specific recommendations for treatment technique.

In Freud's view, abandoned objects form ego structure through the

formation as a matter of the internalization of autonomously moti­

on both the role and content of interpretation and on the extent to

The Work of W. R. D. Fairbairn

Basic Concepts of Fairbairn's Theory

of which the ego is born as well as Hartmann's (1939) concept of an

Pleasure, according to Fairbairn, is a "signpost" to the object. When

Fairbairn was influenced by Klein's (1957) concept of ego development

is complete identification with the object of dependence. Fairbairn

longing. His conceptualization of the second subphase is another

According to Fairbairn, psychic division originates in the infant's

saboteur-rejecting object and the libidinal ego-exciting object while the

relevance for psychopathogenesis. As will be shown later, Fairbairn

Fairbairn's (1941) theory of psychopathology was as developmentally

standing of every form of neurosis, character pathology, and psy­

valuable and therefore to be impoverished. With this formulation,

dependency it becomes anxious that its aggression will destroy the

tion, oral eroticism, masturbation, and other autoerotic activities

In all four types of transitional-phase psychopathology, the patho­

Despite Fairbairn's extensive critique of psychoanalytic theory and his

I wish to express my gratitude to Dr. Paul Stepansky for the excep

relinquishment of these objects, they are taken in through identifica

two sources of ego development: the motivation of inborn appara

group of ego psychologists believes that there has been a "develop

formation as a matter of the internalization of autonomously moti

standing of every form of neurosis, character pathology, and psy

In all four types of transitional-phase psychopathology, the patho

but do not have the capacity is shameful because it exposes an inad

The primary defenses of the paranoid position are projection, intro

adherence to the interpretive method for resolving emotional difficul

Although Klein's writing is replete with evidence of early interpre

The next night the patient dreamed of a pike, to which he associ

projective identifications, but he believed that the result of failed con